Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02262593 2006-11-20
-1-
DATABASE APPARATUS
FIELD OF THE INVENTION
The present invention is in the general field of databases and database
management
systems.
BACKGROUND OF THE INVENTION
As is well known, a database system is a collection of interrelated data files
and a set
of programs that allow one or more users to add data, retrieve and modify the
data stored in
these files. The fundamental concept of a database system is to provide users
with a so
called "abstract" and simplified view of the data (referred to also as data
model or conceptual
structure) which exempts a conventional user from dealing with details such as
how the data
is physically organized and accessed.
Some of the well known data models (i.e. the "Hierarchical model", "Network
model"
and "Relational model" will now be briefly reviewed. A more detailed
discussion can be found
for example in: Henry F. Korth, Abraham Silberschatz, "Database System
Concepts",
McGRAW Hill International Editions, 1986, Chapters3-5 pp.45-172).
Generally speaking, all the models discussed below have a common property in
that
they represent each "entity" as a "record" having one or more "fields" each
being indicative of
a given attribute of the entity (e.g. a record of a given book may have the
following fields
"BOOK ID", "BOOK NAME", "TITLE"). Normally one or more attributes constitute a
"key" i.e. it
uniquely identifies the record. In the latter example "BOOKID" serves as a
key. The various
models are distinguished one from the other, inter alia, in the way that these
records are
organized into a more complex structure.
DOCSMTL: 2234878\1
CA 02262593 1999-O1-26
WO 98/04980 PCT/8,97/00248
-
Relational Model - The relational model, introduced by Cold, is a
landmark in the history of database development. In relational databases an
abstract concept has been introduced, according to which the data is
represented by tables (referred to as "relations") in which the columns
represent the fields and rows represent the records.
The association between tables is only conceptual. It is not a part
of the database definition. Two tables can be implicitly associated by the
fact that they have one or more columns whose values are taken from the
same set of values (called "domain").
Other concepts introduced by the relational model are high level
operators that operate on tables (i.e., both their parameters and results are
tables) and comprehensive data languages (now called 4th generation
languages) in which one specifies what are the required results rather than
how these results are to be produced. Such non-procedural languages (SQL
l~ - Structured Query Language) have become an industry standard. Further-
more, the relational model suggests a very high level of data independence.
There should not be any effect on the programs written in these languages
due to changes in the manner data are organized, stored, indexed and
ordered. The relational model has become a de-facto standard for data
analysts.
Network Model - In the relational model, data (and relationship between
data) are regarded as a collection of tables. In distinction therefrom in the
network model data are represented as a collection of records whereas
relationship between the records (data) are represented as links.
2~ A record in the network model is similar to an "entity" in the
sense that it is a collection of fields each holding one type of data: The
links
may be effectively viewed as pointers. A collection of records and the
relation therebetween constitutes a collection of arbitrary graphs.
Hierarchical Model -The Hierarchical Model resembles the network model
in the manner that data and relations between data are treated, i.e. as
records
and links. However, in distinction from the network model, the records and
the relations between them constitute a collection of trees rather than of
CA 02262593 1999-O1-26
WO 98/04980 PCTIIL97100248
-3-
arbitrary graphs. The structure of the Hierarchical Model is simple and
_ straightforward particularly in the case that the data that needs to be
organized in a database are of inherent hierarchical nature. Consider for
- example a basic entity "Employee" with the following subordinated
attributes "Employee Salary" and "Employee Attendance". The latter may
also have subordinated attributes e.g. "Employee Entries" and "Employ-
ee Exits". In this scenario the data is of inherent hierarchical nature and
therefore should preferably be organized in the hierarchical model. This,
however, is typically not the case. Consider, for example, a scenario where
"Employee" is assigned to several "Projects" and the time he/she spends
("Time Spent") in each project is an attribute that is included in both the
"Employee" and "Projects" entities. Such arrangement of data cannot be
easily organized in the hierarchical model and one possible solution is to
duplicate the item "Time_Spent" and hold it separately in the hierarchies of
I~ "Employee" and "Project". This approach is cumbersome and error prone
in the sense that it is now required to assure that the two instances of
"Time_Spent" are kept identical at all times. Since in real life scenarios
arrangements of data that do not have inherent hierarchial structure are very
common, the hierarchial model is inappropriate for serving as a database in
many real-world scenarios.
As mentioned in the foregoing, data models deal with the
conceptual or logical level of data representation and "hide" details such as
how the data are physically arranged and accessed. The latter characteristics
are normally dealt with by a so-called "database file management system".
2~ The main goal of the database file and system management (referred to
occasionally also as "database engine") is to enhance database performance
_ in terms of time {i.e. from the user's standpoint fast response time of the
database), and space (i.e. to minimize the storage volume that is allocated
' for the database files). As is well known in the art, normally, there is a
trade
off between the time and space requirements. The performance of the
database depends on the efficiency of the data structures that are used to
represent the data and how efficiently the system can operate on these data.
CA 02262593 1999-O1-26
WO 98104980 PCT/IL97/00248
-4-
A detailed discussion on conventional file and management systems is given
for example in Chapters 7 (file system structure) and 8 (indexing and
hashing) in "Database System Concepts", ibid.
A database engine maps the logical structure into physical files
and affords access path to the database records. The following techniques are
typically utilized by known database engines in order facilitate access to
data.
Hashing - This technique is usually a very quick method for locating a
record once the value of its key is known. It involves the translation of the
key into a pointer by some formula and then a direct access. Its drawbacks
are that only one access key can be used on the same record and that a good
translation ("hashing"} formula is not always available. The access usually
requires one I/O operation, but there are cases when the formula maps more
than one record to the same position. This situation requires additional
l~ operations to resolve the conflicts. If the "hashing" formula is good these
cases are rare, and therefore the average number of I/O operations is
somewhat greater but not much greater than one.
Full indexing - This technique can be used to create a virtually unlimited
number of access paths to the same data. The index is a search pattern,
which ultimately locates the data. Its main disadvantages are that it requires
space (usually all the keys to the records plus some pointers) and mainte-
nance (addition and/or deletion of keys whenever a record is added and/or
deleted respectively, or when its key is updated). Normally, the nature of the
indexing technique as well as the volume of the data held in the files
2~ determine the number of I/0 operations that are required to retrieve,
insert,
delete or modify a given data record.
Various types of indexing schemes have been developed but,
normally, an indexing implementation is more costly than the aforemen-
tioned techniques. On the other hand, indexing is the simplest and most
common method for acquiring multiple access paths to the same data. One
of the most widely used indexing algorithms is the B-TREE (under various
commercial product names) in which the keys are kept in a balanced tree
CA 02262593 1999-O1-26
WO 98104980 PCT/IL97/00248
- j
structure and the lowest level points at the data itself. Detailed explanation
of the B+ Tree indexing algorithm (and its derivative indexing algorithm the
B-Tree) can be found in "Database System Concepts" ibid. pp. ?7~-?8?.
The number of I/0 operations obeys the algorithmic expression Log,~N + 1
where K is an implementation dependent constant and N is the total number
of records. This means that the performance slows down exponentially as
the number of records increases, which will at some point, cause unaccept-
ably slow response time.
It is possible, of course, to use a combination of the above or
other techniques, e.g. an indexing technique in combination with a linked
list of records (i.e. records that are serially linked by means of one or bi-
directional pointers). Normally, the beginning of the list (i.e. the first
record
in the list) is acceded by indexing technique and thereafter the pointers are
followed until the sought record is found.
1~ One of the significant drawbacks of the aforementioned popular
B+-Tree indexing algorithm is that the indices portion of the data is not only
held as an integral portion of the data of the leaves of the tree, but is also
held in the interim nodes of the tree serving as a search path for realizing
"FIND", "INSERT" and/or "DELETE" record actions. This results, of
course, in the undesired inflation of the database size and the latter
drawback
is further aggravated when indexes of large size are utilized (i.e. when a
relatively large number of bits is required for representing the index).
One possible approach to cope with this problem is to exploit the
tries (pronounced "try-S") indexing technique discussed, for example, in G.
2~ Wiederhold, "File organization for Database design"; Mcgraw-Hill, 1987,
PP~ 272 273.
Generally speaking, the tries indexing technique enables a rapid
search whilst avoiding the duplication of indexes as manifested for example
by the B+ technique. The tries indexing file has the general structure of a
tree wherein the search is based on partitioning the search according to
search key portions (e.g. search key digit or bit). Thus, for example each
node in the tries indexing file represents a digit position of a search key
and
CA 02262593 1999-O1-26
WO 98/0498U PCT/II,97/00248
-6-
the link to any one of its children represents the digit's value. The tries
structure affords efficient data structure in terms of the memory space that
is allocated therefor, since the search-key is not held, as a whole, in
interim
nodes and hence the duplication that is exhibited for example in the B+
indexing technique is avoided.
In order to achieve enhanced performance in terms of response
time, a tries indexing file should be built by selecting the digits (or bits)
from the search key such that the best possible partition of the search space
in obtained, or in other words so as to accomplish a tree which is as
1Q balanced as possible.
Hitherto known tries indexing file structures have inherent
drawbacks. Thus, for example, as is well known to those versed in the art
and as explained in "File organization for Database design", ibid., the goal
of obtaining a balanced tree necessitates prior knowledge of index values
1~ (which necessarily entails prior knowledge of the data records in the
file).
However, normally there is no prior knowledge of the contents of a database
(e.g. consider a database that holds an inventory of items stored in a
warehouse. Clearly the inventory of items dynamically changes as new
shipments of items are either received from suppliers or delivered to
clients),
?0 the drawbacks of the pre-requisite requirement of knowing the contents of
the database in order to accomplish efficient tries structure are obvious.
Another clear drawback of the conventional tries structure is that the data
is not kept in a sorted form which hinders the conducting of the efficient
search of related items (e.g. this difficulty is exhibited for example when
2~ responding to the query: in a database that holds particulars of supplier,
retrieve the full name and address of all suppliers having a surname that
starts with 'A'). Accordingly, the tries indexing file of the kind specified
exhibits only a theoretical concept which from commercial standpoint is
practically infeasible.
30 It is therefore the object of the present invention to reduce the
drawbacks of data processing systems that exploit hitherto known database
file management system. Specifically, it is the object of the present
CA 02262593 1999-O1-26
WO 98104980 PCT/IL97100248
_'j-
invention to provide for a data processing system that exhibits an enhanced
_ database performance by utilizing an efficient database file management
system.
GLOSSARY
For clarity of explanation, there follows a glossary of terms used
frequently throughout the description and the appended claims. Some of the
terms are conventional and others have been coined.
Block - a storage volume unit which is loaded in its entirety into the
memory of the computer. A block may contain data arranged in any desired
manner, e.g. nodes arranged as a tree and possibly also pointers to actual
data records.
Tree - a data structure consisting of a root node linked by means of n
(n>_1) pointers (or links) to respective n children nodes each of which may
1~ be linked in a similar manner to up to n children nodes and so forth. A
node
in the tree that is not linked to any children node is designated a "leaf
node".
The nodes in the tree that do not constitute "root" or "leaf" are designated
as "interim" nodes. Interim nodes constitute, each, a "root" of a sub-tree of
the specified tree. For any given node in the tree, all the nodes that can be
?0 accessed by moving along the pointers towards the leaves are designated as
"subordinated nodes" of said given node, whereas all the nodes that can be
accessed by moving in the opposite direction, i.e. towards the root are
designated as "predecessor nodes" of said given node.
"Node" should be construed in a broad sense. Thus, the definition
2~ of tree encompasses also a tree of blocks wherein each node constitutes a
block. For detailed definition of "tree", refer also to the book "Graph Algo-
- rithms" by S. Even.
Binary tree - is a tree wherein n=?.
Depth of a tree - is defined as the path in terms of number of nodes that
30 lead from a root node to a given leaf node in the tree.
Balanced tree - a tree having the same depth for all leaf nodes in the tree.
As a rule, a balanced (or essentially balanced) tree normally involves, on the
CA 02262593 1999-O1-26
WO 98/04980 PCT/8.97/00248
-g-
average, less I/O operations in order to access a given data record and is
therefore deemed more efficient than a counterpart non-balanced tree.
File-Indexes (or indexing file) - According to G. Wiederhold, "File
organization for Database design"; Mcgraw-Hill, 1987, pp.I3l, "indexes are
a collection of entries one for each data record, containing a value of a key
attribute for that record, and reference pointer which allows immediate
access to that record". There are known in the art various indexing
techniques such as index-sequential files, B+-Tree or B-Tree index files and
others. The B;-Tree index files comply with the general structure of
essentially balanced trees wherein interim nodes hold index value (or search
keys) of the records and the latter are held at the leaves. As specified
above,
one of the significant shortcomings of B+ Tree structures is that the indexes
are duplicated, i.e. they are held both at the interim level serving for
search
path purposes and at the leaf level as a part of the pertinent data record.
1~ Conventional tries indexing file - The general structure of a conventional
tries structure was defined and discussed in the above "Background of the
Invention" section and is also discussed in the book "File organization for
Database design", ibid.
In the context of the present invention a so called "Probabilistic
?0 Access Indexing File" (referred to also as "PAIF") is utilized which is
distinguished from the conventional tries indexing file, inter alia, in the
following respects:
(i) it does not necessarily comply with the general structure of a
tree, i.e. whereas in a conventional tree each node (except for the root) has
exactl one father node, in the PAIF a node ~ have more than one father;
(ii) whilst the PAIF exploits the general technique of partitioning
the search according to search key portions (e.g. search key digit or search
key bits) as in a conventional tries indexing file), the manner in which this
technique is realized (see below) is well distinguished from the conventional
30 tries indexing file thereby coping with the above referred to inherent
drawbacks of the conventional fife indexing technique and yielding
advantages as will be explained in greater detail below.
CA 02262593 2006-11-20
-9-
It should be noted that for clarity of explanations the hereinbefore "binary",
"balanced"
and "depth" terms that are used with respect to a "tree" also apply to a PAIF.
Thus, for
example, a binary PAIF has at most two children nodes for any given node
therein.
As will be explained in greater detail below, a node in a PAIF is not
necessarily
classified uniquely.
GENERAL DESCRIPTION OF THE INVENTION
There is thus provided in accordance with the invention, a data processing
system
including a processor device associated with memory device; said processor
device is capable
of running at least one user application program capable of communicating with
a database
management system that includes database file management system, for accessing
data
records stored in said memory device; said database file management system
comprising:
at least one probabilistic access indexing file (PAIF) that includes:
a plurality of nodes each of which, except for the leaf nodes, is connected by
means of
at least one link to a respective at least one child node; each leaf node in
said PAIF is
associated with at least one data record accessible to said user application
program; at least a
portion of said data record constitutes an n-long-bit search-key having a most-
significant-bit
(MSB) thereof at an offset 0 and the remaining n-I bits thereof at respective
offsets 1 ton-I;
selected nodes in said PAIF represent, each, a given offset of an I-bit-long
search-key-
portion within said n-long-bit search key and links) originated from each one
of said selected
nodes represents, each, a unique value of said1-bit- long search key portion;
for each one of
said selected nodes, except for the leaf node, there exists at least one node,
subordinated
thereto, having an offset larger than the given offset of said one selected
node;
whereby for any search key in a record associated with a leaf node of said
PAIF, there
is defined a search path consisting of a series of units, each consisting of a
node from among
said selected nodes, and a link, commencing at the root node and ending at
said leaf node
such that for any unit in the series, the value of the I-bit-long search-key-
portion at a given
offset as indicated, respectively by the link and the node of said unit,
conforms the value of the
corresponding I-bit-long portion at said given offset within said search key.
For sake of clarity, a new nomenclature is introduced in order to distinguish
between
two types of links in a PAIF. Thus, "long link" stands for a link which
connects a leaf node to a
DOCSMTL: 223487871
CA 02262593 2006-11-20
- 1~ -
data record, or an inter-block connection, i.e. it connect nodes residing in
two separate blocks.
Long links, typically but not necessarily, necessitate I/O operations. "Short
link" stands for an
intra-block connection, i.e. it links two nodes residing in the same block.
A database file management system that employs a PAIF of the invention is
advantageous, in terms of enhanced performance as compared to hitherto known
techniques
inter alia owing to the following characteristics:
~ The data can be accessed inherently in a sorted form according to a search
key.
~ There is no requirement for in advance knowledge of the contents of the data
base.
~ An essentially balanced tree of blocks is retained and the depth of the tree
is
relatively small, thereby minimizing in the average the number of slow I/O
operations that are required in order to add, delete or retrieve data.
~ Data records of different types (i.e. belonging to different entities) and
search
keys of different lengths may reside in the same PAIF.
~ A search key, as a whole, is held only as an integral part of the data
record
thereby avoiding duplication of search keys, as exhibited for example in the
conventional B+ Tree indexing technique.
A detailed discussion as regards the various advantages offered by the
database file
management system of the invention is given below with reference to specific
embodiments.
It should be noted that the data records may constitute part of the
PAIF, or may reside in one or mode separate data files. In the latter
embodiment the data
records should be linked, of course, to the corresponding PAIF. As will
further be clarified with
reference to the description of specific embodiment below, a given data record
may
accommodate more than one search key.
Still further the invention provides for a data processing system including a
processor
device associated with memory device; said processor device is capable of
running at least
one user application program capable of communicating with a database
management system
that includes database file management system, for accessing data records
stored in said
memory device; said database file management system comprising:
DOCSMTL: 22348780
CA 02262593 2006-11-20
-11-
at least one probabilistic access indexing file (PAIF) having a plurality of
nodes and
links;
the leave nodes of said PAIF are associated each with at least one data record
accessible to said user application program and wherein at least a portion of
said data record
constitutes at least one search-key;
selected nodes in said PAIF represent, each, an offset of a search key portion
within
said search key and links) originated from each given node from among said
selected nodes
represents each a unique value .of said search key portion; for each one of
said selected
nodes, except for the leaf node, there exists at least one node, subordinated
thereto, having an
offset larger than the given offset of said one selected node;
whereby there exists at least one search path in said PAIF commencing from a
root
node and ending at a leaf node that is associated with a given search key such
that the nodes
in said search path are included in said selected nodes; said search path
defines a first series
of i search key portion values at corresponding i offsets which conform to a
second series of i
search key portions at the same i offsets within said given search key.
DOCSMTL: 2234878\1
CA 02262593 1999-O1-26
WO 98/04980 PCTIIL97100248
-1?-
BRIEF DESCRIPTION OF THE DRAWINGS
For a better understanding the invention will now be described,
by way of example only, with reference to the accompanying drawings, in
which:
Fig. 1 shows a generalized block diagram of a system employing a
database file management system;
Fig. 2 shows a sample database structure represented as an Entity
Relationship Diagram (ERD), and serving for illustrative purposes;
Fig. 3 shows the database of Fig. ?, represented as tables in accordance
with the relational data model, with each table holding few data occurrences;
Fig. 4 shows an underlying indexing file structure of the "CLIENT"
table of Fig. 3, in accordance with file management system employing
conventional B+ tree indexing file technique;
Fig. 5 shows an underlying indexing file structure of the "CLIENT"
1~ table of Fig. 3, in accordance with file management system employing
conventional tries indexing file technique;
Figs. 6A-6C show a succession of the underlying indexing file
structure of the "CLIENT" table of Fig. 3, in accordance with file manage-
ment system employing PAIF utilized by a database file management system
?0 of the invention, all residing in a single block;
Figs. 7A-7C show schematic illustrations exemplifying succession of
split block operations, according to one embodiment of the invention;
Fig. 8 shows a single PAiF of a database file management system of
the invention, accommodating data records taken from both the CLIENT and
2~ ACCOUNT tables of the database shown in Fig. 3;
Fig. 9 shows a schematic illustration of a single PAIF utilized by a file
management system according to another embodiment of the invention;
Fig. 10 shows a schematic illustration of a single PAIF utilized by a
file management system according to yet another embodiment of the
30 invention;
CA 02262593 1999-O1-26
WO 98104980 PCT/8.97/00248
-13-
Fig. 11 shows a schematic illustration of a single PAIF utilized by a
_ file management system according to still yet another embodiment of the
mventton;
Fig. 12A-D show four benchmark graphs demonstrating the enhanced
performance, in terms of response time and file size, of a database utilizing
a file management system of the invention vs. commercially available Ctree
based database; and
Fig. 13A-D show four benchmark graphs demonstrating the enhanced
performance, in terms of response time and file size, of a database utilizing
a file management system of the invention vs. commercially available Btree
based database.
DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
Attention is first directed to Fig. 1 showing a generalized block
1~ diagram of a system employing a database file management system of the
invention. Thus, a general purpose computer 1, e.g. a personal computer
(P.C.) employing a Pentiurri microprocessor 3 commercially available from
Intel Corp. U.S.A, has an operating system module ~, e.g. Windows NT'
commercially available from Microsoft Inc. U.S.A., which communicates
with processor 3 and controls the overall operation of computer 1.
P.C. 1 further accommodates a plurality of user application
programs of which only three 7, 9 and 11, respectively are shown. The user
application programs are executed by processor 3 under the control of
operating system ~, in a known per se manner, and are responsive to user
2~ input fed through keyboard 13 by the intermediary of I/O port l~ and the
operating system ~. The user application programs further communicate
- with monitor 16 for displaying data, by the intermediary of I/O port 17 and
operating system ~. The user application programs can access data stored
in a database by means of database management system module ?0. The
generalized database management system, as depicted generally in fig. i,
includes high level management system ?? which views, as a rule, the
underlying data in a "logical" manner and is responsive, to the user
CA 02262593 1999-O1-26
WO 98/04980 PCT/8.97/00248
-14-
application program by means known per se such as, e.g., SQL Data
Definition and Data Manipulation language (DDL and DML). The database
management system typically exploits, in a known per se manner, a data
dictionary ?4 which describes the logical structure of the underlying data.
The underlying structure of the data is governed by database file
management system 26 which is associated with the actual data records 28.
The "high-level" logical instructions (e.g. SQL commands) received and
processed by the high-level management system 22 are converted into
"lower level" commands that facilitate access paths to the data records that
are stored in the database files) and to this end the database file manage-
ment system considers the actual structure and organization of the data
records. The -"high level" and "low level" portions of the database file
management system can communicate through a known per se Application
Programmers Interface (API), e.g. the ODBC standard ver. 2.0 commercially
l~ available from Microsoft. The utilization of the ODBC enables "high level"
modules of the database file management system to transparently communi-
cate with different "database engines" that support the ODBC standard. As
is well known to those versed in the art, the advantage of the latter
characteristics is brought about, for example, by substituting a new, faster
''0 and more efficient ODBC compatible "database engine" for an older one,
whilst retaining the "higher level" modules of the database management
system intact. The term access to data or data records used herein
encompass all kind of data manipulation including "find", "insert", "delete"
and "modify" data record(s), and the pertinent DDL commands which afford
2~ the construction, modification and deletion of the database.
Fig. 1 further shows, schematically, an internal memory module
?9 (e.g. 16 Mega byte and possibly employing a cache memory sub-
module) and an external memory module 29' (e.g. 1 gigabyte). Typically,
external memory ''9 is accessed through an external, relatively slow
30 communication bus (not shown), whereas the internal memory is normally
accessed by means of a faster internal bus (not shown). Normally, by virtue
of the relatively small size of the internal memory, only those applications
CA 02262593 1999-O1-26
WO 98104980 PCT/8.97100248
-1J-
(or portions thereof) tnat are currently executed are loaded from the external
memory into the internal memory. By the same token, for large databases
that cannot be accommodated in their entirety in the internal memory, a
major portion thereof is stored in the external memory. Thus, in response
to an application generated query that seeks for one or more data records in
the database, the database management system utilizes operating system
services (i.e. an I/O operation) in order to load, through the external
communication bus, one or more blocks of data from the external to the
internal memory. If the sought data records are not found in the loaded
blocks, successive I/O operations are required until the sought data records
are targeted.
It should be noted that for simplicity of presentation, the internal
and external memory modules ?9, ?9', are separated from the various
modules s, 7, 9, 11, ?0. Clearly, albeit not shown, the various modules
l~ (operating system, DBMS, and user application programs) are normally
stored in the external memory and their currently executed portions are
loaded to the internal memory.
Computer 1 may serve as a workstation forming part of a LAN
Local Area Network (LAN) (not shown) which employs a server having also
essentially the same structure of Fig. 1. To the extent that the workstations
and the server employ client-server based protocols a predominant portion
of said modules (including the database file management system ?6 and the
database records themselves ?8) reside in the server.
Those versed in the art will readily appreciate that the foregoing
2~ embodiments described with reference to of Fig. 1 are only two out of many
possible variants. Thus, by way of non-limiting example, the database may
be an on-line database residing in an Internet Web site. It should be further
noted that for clarity of explanation system 1 is illustrated in a simplified
and generalized manner. A more detailed discussion of database file
management systems and in particular of the various components that are
normally accommodated in database file management systems can be found,
e.g. in Chapter 7 of "Database System Concepts" ibid.
CA 02262593 1999-O1-26
WO 98/04980 PCTlIL97/00248
-16-
Having described the general structure of a system of the
invention, attention is now directed to Fig. 2 showing a sample database
structure represented as Entity Relationship Diagram (ERD), and serving for
illustrative purposes. Thus, the ERD 30 of Fig. ? consists of the entities
a "CLIENT" 32 and "ACCOUNT" 34 as well as an "n to m" "DEPOSIT" 36
relationship indicating that a given client may have more than one account
and by the same token a given account may be owned by more than one
client.
As shown, the entity "CLIENT" has the following attributes
(fields): "Client-Id" 38 being a key attribute that uniquely identifies each
client, "Name" 39 standing for the client's name and "Address" 40 standing
for the client's address. The entity "ACCOUNT" has the following attributes
(fields): "Acc No" 4? being a key attribute that uniquely identifies each
account, and "Balance" 43 holding the balance of the account. The
1~ relationship "DEPOSIT" consists of pairs of keys of the "CLIENT" and
"ACCOUNT" entities, such that each pair is indicative of particular account
owned by specific client.
Turning now to Fig. 3, there is shown a database of Fig. ?,
represented as three tables ~0, ~1 and ~2 corresponding to the relational data
model, 32, 34 and 36, respectively, with each table holding a few data
occurrences for illustrative purposes. It should be noted that the length of
the key field ("Client ID") of the "CLIENT" table is ~ digits, whereas the
length of the key field ("Acc ID") of the "ACCOUNT" table is 6 digits.
The client table holds ~ data occurrences ~~-~9, the account table holds ?
2~ data occurrences 6~, 66 and the deposit table holds 3 data
occurrences 70-72.
in accordance with prior art techniques each table is, as a rule,
organized in a different file. Thus, Fig. 4 illustrates an underlying indexing
file structure of the "CLIENT" table of Fig. 3, in accordance with file
management system employing the conventional B+ tree indexing file
technique. As shown, the indexing file 80 consists of three blocks 80a-c
standing for a root block and two leaf blocks respectively. The data records
CA 02262593 2006-11-20
-17-
are organized in a separate file 81 holding the five data records 55-59 (in
Fig. 3). Each
block consists of a succession of pair of fields (e.g. 82a-b and 83a-b in
block 80a). In each
pair the first field stands for a search key value and the second field stands
for a pointer to
the next block to search, or in the case of a leaf block an address of the
data record.
Thus, block search key12355 (82a) in block 80a can be found in block 80b (by
pointer 82b). In block 80b, the search key 12355 (86a) is associated with
pointer "5" (86b)
indicating the address of the data record identified by this search key in the
data file 81. Put
differently the data record that is identified by search key "12355" (57 in
Fig. 3) is the fifth in
order in data file 81.
The tables "ACCOUNT" and "DEPOSIT" are likewise arranged in two separate tries
tree indexing files, respectively.
The B+ tree indexing file of Fig. 4 exhibits one of the significant
shortcomings of this
approach in that the keys (i.e. search keys) are duplicated, i,e. they are
held both in the
interim nodes and in the data file associated with the leaves of the B+ tree
indexing file.
Thus, for example, the search key of data record 57 is not only held as an
integral part of
the data record in file 81 but also in block 80a (search key 82a) and in block
80b (search
key 86a).
This being the case, one readily notices that for large files (which is the
case in
many real-life scenarios) the duplication of the search keys (and particularly
where n-long-
key for large n is concerned) results in inflated indexing files) which
necessitate a large
storage volume.
Fig.5 illustrates an underlying indexing file structure of the "CLIENT" table
of Fig. 3,
in accordance with a file management system employing the conventional tries
indexing file
technique. Thus, tries indexing file 90 consists of plurality of nodes and
links wherein each
node stands for a position and the link stands for a value at this position.
Table 91 has four
columns. The first column indicates which digit position is to be used, the
second column
the value of that digit. A digit value partitions the key into two subsets.
Column three and
four direct the search procedure to the next step.
In order to locate a given search key, e.g.12355, a digit at the position
indicated by
the root (position "5" indicated by node 90a, being also the first column in
the first line of
DOCSMTL: 2234878\1
CA 02262593 2006-11-20
-18-
table 91 ) is compared to the value specified at the second column of the same
line (value
"5" indicated also by link 90b in the tries structure). Since the digit at
position 5 of the
sought search key 12355 is indeed 5, control is transferred to line 2 (as
indicated by the
third column of line 1 of table 91 ). Next, the digit at position 3 of the
sought search key (90c
in the tree, being also the value of the first column of the second line in
table 91 ) is
compared to the value 3 (link 90d, being also the second column in the second
line of the
table 91 ). Since match occurs, control is transferred to line 3 in the table.
In this step the
digit at position 4 of the sought search key does not match the value
specified at the
second column of line three (i.e."5" vs. "4") and accordingly as indicated in
the fourth
column of table 91 ("not equal") a pointer to the sought data record (57) is
obtained.
The tables "ACCOUNT" and "DEPOSIT" are likewise arranged in two separate tries
indexing files, respectively. In contrast to the B+ tree indexing file of Fig.
4, the one shown
Fig.5 does not necessitate duplication of the search key. Put differently,
only one instance
of each search key is held, as a whole, in the tries indexing file. In this
sense it constitutes
an advantage over the B+ technique.
However, and as specified above, in order to retain an essentially balanced
tree
prior knowledge of the database contents is required. This requirement poses
undue
constraint since databases of the kind described in Fig.2 are of a dynamic
nature, e.g. for
the specific database of Fig. 2, new clients open accounts, senior clients
close accounts,
new clients are registered as co-owners of existing accounts etc. Each of
these
transactions, when applied to the conventional tries indexing file may require
substantive
rearrangement of the tries structure (which is obviously undesired) in order
to retain the
"balanced tree" characteristics.
DOCSMTL: 2234A7A\I
CA 02262593 1999-O1-26
WO 98104980 PCTIIL97/00248
-19-
Having shown the file indexing structure according to two
hitherto known techniques, attention is now directed to Figs. 6A-6C
showing a succession of underlying indexing file structures of the "CLIENT"
table of Fig. 3, in accordance with the file management system employing
the PAIF of the invention. The terms "transaction" and "operation" are used
interchangeably.
In the description below the basic commands which enable data
manipulation in the PAIF will be reviewed, i.e. insert new data record to an
PAIF, find data record in PAIF, and delete existing data record. Those
versed in the art will no doubt appreciate that on the basis of these basic
primitives more compound data manipulation operations, (e.g. "Join") may
be realized.
Turning at the onset to Fig. 6a, there is shown the Client's data
record 103 (~6 in table Client of Fig. 3) having search key "1?34~" (i.e. a
1~ ~-long-digit search key). The PAIF of Fig. 6a (100) is, of course, trivial
and consists of a single node 101 (standing for both the root node and the
leaf node) linked by means of a long link 10? to data record 103.
The node 100 represents an offset 0 in said search key and the
link 10'? represents a value "1" of the search key portion (being by this
?0 particular embodiment 1-digit-long) at the specified offset.
As clearly shown in Fig. 6a, the data record 103 is associated
with a search path being a unit that consists of a node 101 and a link 102
which defines an offset and a pertinent search key portion value that
conforms to the corresponding search key portion value at that particular
offset within the search key of the specified data record. More specifically,
the value of the one-digit search-key-portion at offset 0 within search key
"1?34~" is indeed "1".
Turning now to Fig. 6b there is shown a PAIF 108 after the
' termination of a successive transaction in which the data record having
30 Client Id No "I244~" 107 has been inserted (data occurrence 58 in table
Client of Fig. 3). The search keys of data records 103 and I07 are
distinguished only in the third digit (offset ?), i.e. "3" and "4"
respectively.
CA 02262593 2006-11-20
-20-
The unit defined by root. node 101 and the link 102 is not sufficient to
discern
between data records 103 and 107, since the value of the 1-digit search key
portion at offset
0 for both data records is "1". Hence, node 104 indicates the lowest offset
which
distinguishes between the two records and short links 105 and 106 indicate on
the
respective 1-digit search key portion "3" and "4" at offset 2.
Before moving on to describe a procedure of inserting a new data record to an
existing database it should be borne in mind that the higher the node in the
PAIF the smaller
is the offset indicated thereby (e.g. in the PAIF of Fig. 6B, node 101 is
higher than mode 104
and accordingly it is assigned with smaller offset --"0" vs."2").
Generally speaking, the preferred procedure for inserting a new data record
into an
existing PAIF includes the execution of the following steps:
i. advancing along a reference path commencing from the root node and ending
at
a data record associated to a leaf node (referred to as "reference data
record");
ii. comparing the search key of the reference data record to that of the new
data
record for determining the smallest offset of the search key portion that
discerns
the two (hereinafter discerning offset).
iii. proceed to one of the following steps (ii.0-iii.3) depending upon the
value of the discerning offset:
iii.0 if the data records are equal then terminate; or
iii.l if the discerning offset matches the offset indicated by one of the
nodes in the
reference path, add another link originating from said one node and assign to
said
link the value of the search key portion at the discerning offset taken from
the search
key of the new data record; or
iii.2 if the discerning offset is larger than that indicated by the leaf node
that is
linked, by means of a link, to the reference data record:
DOCSMTL: 2234878\1
CA 02262593 1999-O1-26
WO 98/04980 PCT/8,97/00248
- ~1 -
iii.?.1 disconnect the link from the reference data record (i.e.
it remains temporarily "loose") and move the link to a
new node; the new node is assigned with a value of
the discerning offset;
iii.?.? connect the reference data record to the new node
(which now becomes a leaf node) and assign to the link
(long link) a value of the search-key-portion at the
discerning offset taken from the search key of the
reference data record;
iii.?.3 connect by means of a link the new data record to the
new node and assign to the link (long link) a value of
the search-key-portion at the discerning offset taken
from the search key of the new data record; or
iii.3 if conditions iii.0,iii.1 and iii.? are not met, there exists, in
1~ the reference search path, a father node and a child node
thereof such that the discerning offset is, at the same time,
larger than the offset assigned to the father node and smaller
than the offset assigned to the child node. Accordingly, apply
the following sub-steps:
iii.3.1 disconnect the link from the father node to the child
node and shift the link to a new interim node (i.e. the
child node remains temporarily "loose"); The new
interim node is assigned with the value of said dis-
cerning offset;
iii.3.? connect by means of a link (short link) the new data
record to said new interim node; the value assigned to
the link is that of the search-key-portion at the dis-
cerning offset, as taken from the search key of the new
data record;
iii.3.3 connect by means of a new link the new interim node
to the child node (i.e. the new interim node becomes a
new father node thereof), and the value assigned to said
CA 02262593 2006-11-20
-22-
link is the search-key-portion at the offset indicated by the new father node,
taken from the search key of the reference data record.
It should be noted that for a different reference path a different PAIF may be
obtained.
For a better understanding, the aforementioned "insert data record" operation
will be
successively applied to the specific PAIF of Fig. 6B, each time with a
different data record so
as to exemplify the three distinct scenarios stipulated in steps iii.1 -
iii.3. above, thereby
resulting in three PAIF illustrated in Figs. 6C-1 to 6C-3, respectively.
In the first example the CLIENT data record having Client_Id (or search key)
"12546"
(59 in table Client of Fig. 3) is inserted to the PAIF of Fig. 6B. As
stipulated in step (i) we
move along the reference path commencing from the root 101 to node 104 (see
"Find data
record" defined below) and from there, the "Not Found" state by an arbitrary
path that ends,
for example, at data record 103 which stands for the "reference data record".
The
comparison operation stipulated in step (ii) results in that the search key of
the new data
record in distinguished from the search key of the reference data record at
offsets 2 ("5" vs.
"3") and 4 ("6" vs. "5"). The smallest offset ("discerning offset") is
therefore 2.
Turning now to step (iii), the condition of step iii.1 is met since the
discerning offset is
equal to that assigned to node 104. Accordingly, and as is shown in Fig. 6C-1,
new link 111
connects node 104 to the new data record112. The value assigned to link 111 is
5, being
the digit at location2 in the search key of the new data record112. PAIF 110
of Fig. 6C-1 is
therefore the result of inserting the data record 112 into the PAIF 108 of
Fig. 6B.
Moving now to the second example, the CLIENT data record having Client Id (or
search key) "12355" (57 in table Client of Fig. 3) is inserted into the PAIF
of Fig. 6B. Steps i
and ii, stipulated above result for the same reference path in the discerning
offset 3 (since
the only distinction between the search key of the new data record and that of
record 103 is
in the digit at location 3 - "5" vs. "4").
DOCSMTL: 2234878\1
CA 02262593 1999-O1-26
WO 98/04980 PCT/IL97/00248
-23-
Turning now to step (iii), the condition of step iii.2 is met
since the discerning offset 3 is larger than the offset ? of leaf node 104 in
the reference search path. Accordingly, in compliance with step iii.2.l and
as is shown in the resulting PAIF 120 of Fig. 6C-?, the link 106 is
disconnected from reference data record 103 and is connected to a new node
121. The new node is assigned with the discerning offset 3. Next, in
compliance with step iii.2.2, the reference data .record 103 is connected to
the new node 121 by means of new link 122. The new link is assigned with
the value 4 (being the digit at the discerning offset 3 taken from the search
key "1?34~" of the reference data record 103); and finally, as stipulated in
step iii.2.3, the new data record 123 is connected to node 121 by means of
link 124 which is assigned with the value "~" (being the digit at the
discerning offset 3 taken from the search key "1?3~~" of the new data
record 123). PAIF 120 of Fig. 6C-? is, therefore, the result of inserting the
l~ data record 123 into the PAIF 108 of Fig. 6B.
The third and last example concerns inserting the CLIENT data
record having Client Id (or._, search key) "11346" (~~ in table Client of
Fig. 3) into the PAIF of Fig. 6B. Applying the aforementioned steps i and
ii result in discerning offset 1.
Thus in step iii, the condition of step iii.3 is met. Accordingly,
in compliance with step iii.3.1 and as is shown in the resulting PAIF 130 of
Fig. 6C-3, the link 102 is shifted to a new interim node 131. The new
interim node 131 is assigned with the value 1 (being the discerning offset).
As stipulated in step iii.3.2, the new data record 132 is directly connected
by means of new link 133 to node 131; the value assigned to link 133 is 1
(being the digit at the discerning offset 1 taken from the search key "11346"
of the new data record 132), and finally, in compliance with step iii.3.3 the
new interim node 131 is linked to node 104 by means of link 134 assigned
with the value 2 ((being the digit at the discerning offset (1) taken from the
search key "1234" of the reference data record 103).
Although the PAIF described above with reference to Fig. 6A-
6C may be accommodated within one block it is nevertheless preferable to
CA 02262593 2006-11-20
-24-
separate between "nodes" and "data records" such that data records are grouped
in a
distinct file or files. Applying this approach to the PAIF of Fig. 6C-3,
results in the generation
of the data record file holding the records 103, 107 and 132. Links 100, 105
and 133
become, of course, long links.
Obviously, if an insert procedure results in finding that the data record to
be inserted
already exists in the PAIF an appropriate error message is returned to the
procedure that
invoked the Insert command.
It should be noted that in the latter examples it is assumed that the entire
PAIF
resides in a single block. Obviously when additional data records are inserted
by following
the foregoing "insert procedure" a block overflow may occur, which
necessitates invoking
"split block" procedure as described below.
Having described a typical "Insert" operation, a "Find (or
Retrieve) data record" procedure will be now described. Thus, for finding a
data record
having a given search key (hereinafter the sought data record) in an existing
PAIF. the
following steps should be executed:
i. advance along a search path commencing from the root node and ending at
a data record linked to a leaf node, and for each node in the search path
(hereinafter
"current node") perform the following sub-steps:
i.1 for each link originated from the current node: compare the search-
key-portion of the sought data record at the offset defined by the current
node
to a search-key portion value assigned to said link; in case of a match
advance along said link and return to step i;
i.2. if none of the links originated from the current node matches the
search-key-portion of the sought data record, return "NOT FOUND" and
terminate the find procedure;
i.3 if a data record is reached (hereinafter "reference data record"),
compare the search key of the sought data record as a whole, to that of the
reference data record;
DOCSMTL: 2234878\I
CA 02262593 1999-O1-26
WO 98104980 PCTIB.97100248
_ '7 j _
i.3.1 in case of a match, return "FOUND" (and
in case of "Retrieve", return also the
entire data record) and terminate the find
procedure; or
i.3.? in the case of mismatch return "NOT
FOUND" and terminate the find proce-
dure.
For a better understanding the "find" procedure will be applied,
twice, to the specific PAIF of Fig. 6C-3 giving rise to "found" and "not
found", results respectively.
Thus, consider a find data record having search key "1''44"
(herein after sought data record). According to step i.1 the value of the
digit
"1" at the offset assigned to the root node {offset 0) of the sought data
record is compared to the one assigned to link 10? {being the sole link
1~ originated from node 101). Since a match is found, control is shifted to
node
131. Again according to step i.l the value of the digit ("?") at the offset
assigned to node 131 (offset 1) of the sought data record is compared to the
one assigned to link 134 {being the sole link originated from node 101).
Here also a match is found so control is shifted to node 104. Next,
according to step i.l, the value of the digit "4" at the offset assigned to
node 104 (offset ?) of the sought data record is compared for each link
originating from mode 104. The comparison results in a match for link I0~
and accordingly control is shifted to data record 107.
According to step i.3 the search key of the sought data record
and that of data record 107 are compared and since a match is found a
"FOUND" result is returned (step i.3.1).
Turning now to a second example, consider the case when the
sought data record has a search key "1?463". The procedure described with
reference to the previous example is repeated, however at step i.3 the
comparison between the sought data record and data record 107 results in
a mismatch, and according to step i.3.? a "NOT FOUND" result is returned.
CA 02262593 1999-O1-26
WO 98/04980 PCTlIL97100248
-?6-
A general "Delete Data Record" procedure will now be
described. Thus, as a first stage a "Find data record" procedure is applied to
the PAIF. In case of "NO FOUND", an appropriate error message is
returned to the procedure that invoked the "Delete" command. Alternatively,
the sought data record is found. For clarity of explanation of the "Delete"
procedure, the following nomenclatures are introduced:
The leaf node that is linked to the sought data record is
referred to as the "target node". The father of the target node is referred to
as the "predecessor target node". The link that connects the predecessor
IO target node to the target node is referred to as the "predecessor link" and
the
link that connects the target node to a child node thereof (or to a data
record
other than the sought data record) is referred to as the "target link".
Bearing this nomenclature in mind, the following steps are executed:
i. delete the sought data record and the link that links the
I~ target node thereto;
ii. if the number of links that remain in the target node is
larger than or equal to ?, then the deletion procedure
terminates;
iii. if, on the other hand, the number of links that remain
20 in the target node is exactly one (i.e. one target link),
then:
iii.l "bypass" the target node by connecting the
predecessor link from the predecessor node to
said child node (or to a data record); and
2~ iii.? delete the target node and the target link; termi-
nating the deletion procedure. It should be noted
that the current step is more of "prudent memory
management" step in order to release the space
occupied by the target node and link, so as to
30 enable allocation thereof to other nodes and
links in the block.
CA 02262593 1999-O1-26
WO 98104980 PCT/IL97/00248
For a hetter understanding the foregoing "delete data record"
procedure will be applied to the specific PAIF of Fig. 6C-3.
Thus, responsive to a command "delete record having search
key = "11346", the latter record is searched in the PAIF according to the
procedure described above. Having found the data record 132 and in
compliance with step i above, the data record as well as the link 133 leading
thereto are both deleted. Since after the latter deleting step, the target
node 131 remains only with the sole target link 134, step iii and iii.l apply,
and accordingly the predecessor link 10? bypasses target node 131 and is
directly linked to the child node thereof 104. Next, in compliance with
step ii.?, target node 131 and the target link 134 are deleted thereby
obtaining the PAIF shown in Fig. 6B.
Another Example is given with reference to the PAIF of
Fig. 6C-1. Thus, responsive to a command "delete record having search key
- "12546", the latter record is searched in the PAIF according to the
procedure described above. Having found the data record 11? and in
compliance with step i above, the data record as well as the link (111)
leading thereto are both deleted. Since, as stipulated in step ii, the number
of links that remain in the target node 104 is two (i.e. links 105 and 106),
then the deletion procedure terminates. The resulting PAIF is again the one
shown in Fig. 6B.
Another common primitive is the "Modify existing data
record", e.g. change the address of an existing client. The "Modify"
primitive is normally realized by selectively utilizing the aforementioned
primitives. For executing a "Modify" command one should distinguish
between the following cases:
1. The "modify" applies to fields other than the search key
(e.g. modify the address of a client having Client_Id_No ="xxxxx") - in this
case the modify procedure simply involves a "Find" operation {data record
having Client_Id_No ="xxxxx"). Having found the sought data record, the
old address is replaced by a new one.
CA 02262593 2006-11-20
-28-
2. The "modify" applies to a search key field (e.g. change an account no. from
"xxxxxx" to"yyyyyy"). This command is realized as a sequence of two other
primitives, i.e.
delete data record having "Account No" ="xxxxxx" and thereafter insert data
record having
"Account No" ="yyyyyy", or vice versa. Obviously a Modify transaction may
consist of both
cases.
In the previous examples each search key is represented as a series of digits
and
accordingly the search procedure is performed by partitioning the search-key
into search
key portions each consisting of at least one digit.
Those versed in the art will readily appreciate that digits are not the only
possible
representation of a search key. Thus, for example, a search key can be
represented in
binary form, i.e. a series of 1's and 0's and accordingly the search procedure
is performed
by partitioning the search-key into search key portions each consisting of one
bit (i.e. I=1 ) on
byte (i.e. I=8 bits) and others. In certain scenarios, it may well be the case
that the I value is
not identical for all the nodes in the PAIF.
It should be further noted that different links in a given PAID may be
assigned with
search-key-portions of different length.
As is clearly evident from various PAIF of Figs. 6a-6c, the data records can
be
accessed in a sorted form according to search key. Scanning, for example, the
PAIF of
Figs. 63-C (from left to right) brings about the ordered series "11346",
"12345", and "12445".
This characteristics constitutes yet another advantage which ease data
manipulation. As
specified before, a node in the PAID is not necessarily classified uniquely.
Thus, for
example, in the PAIF 120 of Fig. 6C-2, node 104 is at the same time a lead
node (linked, by
means of a long link 105 to data record 107) and an interim node (linked by
means of a
short link 106 to node 121 ).
Those versed in the art will readily understand that the "Insert", "Delete",
"Find", and
"Modify" procedures described herein are only one out of many possible
variants for
realizing these procedures and they may be modified, all as required and
appropriate
depending upon the particular application.
Turning now to Fig. 7A-C, there are shown schematic illustrations exemplifying
succession of split block operations, according to one embodiment of the
invention.
DOCSMTL: 2234878U
CA 02262593 2006-11-20
-29-
Consider for example a block140 which is unable to accommodate (in terms of
memory
space) the sub-PAIF 141. This being the case a "split block" procedure is
invoked which
results in the block tree structure 142 consisting of root block 144 linked to
leaf block 146 by
means of direct link 145 (as opposed to "short link" and "long link" specified
above) and by
means of long link 147 to a leaf block 148. By this specific example, the
split point was
selected to be link 149 (hereinafter "split link") thereby shifting nodes A,
B, E D and F to new
block 146 and nodes C, G, I, J, K, L and H to a new block 148. The split link
is preferably
selected in order to accomplish an essentially even distribution of nodes and
links between
the new blocks (e.g. the size of the sub PAIFs that resides in blocks 148 and
149 is
essentially the same). The node from which the split link is originated is
normally duplicated
and reside in the father block (144) and the connection between this node and
the replicated
copy thereof in block 146 is implemented by means of said direct link. The
link 149 (being
originally a short link) is replaced by long link 147. Optionally node A and C
may also be
linked by means of another long link marked as dashed line 150.
If, for sake of discussion, the sub PAIF cannot be accommodated in block 148
it
undergoes similar block split procedure resulting in block tree structure 151
in Fig. 7C. By
this example the split link is short link152 and accordingly nodes G, I, K, J
and L are
arranged in a new leaf block 144 and node C is copied to the latter. The node
from which
the split link originates (node C -153) is replicated (yielding an additional
node 153a) and
placed both in blocks 140 and 148A. As before, nodes 153 and153a are linked by
means of
direct link (154). As is clearly seen in Figs. 7B and 7C split block procedure
results in a
balanced tree of blocks thereby keeping the tree depth to a minimum and
consequently
minimizing
DOCSMTL: 2234878\1
CA 02262593 1999-O1-26
WO 98/04980 PCT/a.97/00248
-30-
the number of I/O operations that are required it order to find, insert or
delete a given data record.
The resulting structure does not adversely affect the character-
istics of the general purpose PAIF described above. Thus, for example,
reverting to block 140 (Fig. 7A) it is readily seen that the search path to
node L consists of the following nodes A,C,G,I and L. As shown in Fig. 7C,
the counterpart search path being A,C (block 140), G,I and L (block 148B).
By another example the search path to node H consists of nodes A,C and H
in block 140 whereas the counterpart search path in Fig. 7C being A,C,C,
and H bringing about only limited overhead (i.e. "C,C" in the latter search
path vs. "C" in the former search path). The replicated node and its
associated direct pointer are added for the purpose of managing efficiently
a PAIF which extends to plurality of blocks, and are therefore not deemed
as selected nodes in the sense defined above. Put differently, by this
1~ particular embodiment the number of selected nodes in the PAIF do not
match the total number of nodes in the PAIF.
The opposite procedure "Delete block" is activated when a data
record is deleted leaving only one node in a block having no data records
associated therewith.
Those versed in the art will readily understand that the "split
block" procedure described herein is only one out of many possible variants
for realize the "split block" procedure, so as to retain the PAIF characteris-
tics and accomplish an essentially balanced tree of blocks.
As discussed in detail above, a database management system
that exploits file management system based on conventional tries structure
offers advantages in terms of reduced file size as compared to counterpart
systems that utilize, for example, commercially available B+ tree indexing
files. These advantages stem predominantly from the partitioning of the
search according to search key portions. Since a file management system
that employs the PAIF of the invention also utilizes partitioning the search
according to search key portions, it likewise benefits from the same
advantages.
CA 02262593 1999-O1-26
WO 98/04980 PCTI)297/00248
-31 -
However, and as discussed iZ detail in the foregoing, the
conventional tries indexing file only purports to offer an improved solution
since it poses undue constraints such as the requirement of prior knowledge
on the contents of the database, and others. These constraints render a
solution based on conventional tries indexing file practically infeasible from
commercial standpoint.
A database file management system of the invention not only
copes with the drawbacks of the conventional tries indexing file but also
offers other benefits which facilitate and improve data access by user
application programs.
Thus, the fact that an essentially balanced tree of blocks is
retained assures that, on the average, the number of slow I/O operations is
retained essentially optimal.
It should be emphasized that whilst according to the PAIF of
l~ the invention an essentially balanced tree of blocks is retained, the
latter
characteristics do not necessarily apply to the sub-probabilistic accesses
indexing file (sub PAIF) structure accommodated within a given block. Put
differently, within a given block that is loaded to the memory the sub PAIF
may be unbalanced (as exhibited for example by the PAIF 1?0 of
Fig. 6C-2, wherein the depth of data record 107 is two and the depth of
data records 103 and 1?3 is three}.
Whilst the fact that a sub-PAIF that resides within a block
may be unbalanced is seemingly a drawback, those versed in the art will
readily appreciate that its implications on the overall database performance
2~ are virtually insignificant, since the search operations performed on the
unbalanced sub-PAIF are all executed within a block i.e. in the fast internal
memory of the computer system. As opposed to the intra-block structure,
which may optionally be unbalanced, the arrangement of a block within a
tree is retained essentially balanced thereby minimizing, on the average, the
number of I/O accesses to the external memory (an operation which is
inherently slow) in order to load a desired block to the internal memory.
CA 02262593 2006-11-20
-32-
As is clearly evident from the examples in Figs. 6a- 6c an PAIF of the
invention may
be constructed incrementally and in contrast to conventional Tries indexing
file, no prior
knowledge of the database contents (i.e. the data records) is needed.
A further advantage brought about by utilizing the database file management
system
of the invention is that data is inherently held in a sorted form according to
search key. The
latter characteristic stems from the fact that for any given node in the PAIF,
the series of
search key portions (that are defined by the search path commencing at the
root node and
ending at said given node) appear at each one of the search keys of the data
records (and at
the same offsets) that are associated with the leaf nodes that are
subordinated to said given
node.
For a better understanding of the foregoing, attention is drawn to Fig. 6C-2.
Focusing
for example on node 104, the pertinent search path includes node 101 and
Iink102 signifying
a value "1" at offset 0. As expected, all the data records that are
subordinated to node 104,
i.e. data records107,123 and 103 meet this provision, or in other words the
value of the digit
at offset 0 in each one of the search keys of specified data records is indeed
"1"
Focusing, for example, on node 121, the pertinent search path includes node
101, link
102, node 104 and link 103, signifying a value "1" at offset 0 and value "3"
at offset 2. As
expected, all the data records that are subordinated to node 121, i.e. data
records 123 and
103 meet this provision, or in other words, the value of the digit at offset 0
in both records is
indeed "1" and likewise the value of the digit at offset 2 in both records is
indeed "3". In
contrast, data record 107, which is not subordinated to node 121 does not
comply with the
latter provision, since the value of the digit at offset 2 thereof is "4".
Accordingly, data records having "neighboring" search keys reside in the same
cluster
of the PAIF. The latter characteristics facilitates access to data by user
application programs.
As specified in a PAIF of the invention each interim node holds only a portion
of a
search key whereas in commercially available file management systems based for
example
on B+ tree indexing file arrangement, an interim node holds two or more search
keys in their
entirety. Hence, the memory space that should be allocated for holding an
interim node in a
PAIF is significantly smaller than that required to hold a counterpart node in
the B+ tree
DOCSMTL: 2234878U
CA 02262593 2006-11-20
-33-
indexing file. Based on the foregoing observation, it readily arises that a
fixed size block that
is loaded in the memory accommodates more PAIF interim nodes and their
pertinent links
than a counterpart block accommodating B+ interim nodes. Accordingly the sum-
total of
blocks required to represent a given database in a file management system of
the invention is
smaller than the number of blocks required to represent the same database,
based on B+
tree indexing file. Bearing further in mind that the file management system of
the invention
assures an essentially balanced tree of blocks, the depth of the PAIF that is
required to
represent a given database is smaller than the counterpart indexing file
according to B+
technique.
Accordingly, a database file management system that utilizes a PAIF of the
invention
affords a reduced number of I/O operations, on the average, in order to access
data, thereby
further enhancing database performance.
As stated before, the database file management system of the invention enables
to
hold different types of data in one PAIF. Thus, for example, data records that
form instances
of the CLIENT and ACCOUNT entities can both reside in the same PAIF. Fig. 8
illustrates a
PAIF 160 consisting of the PAIF 130 of Fig. 6C-3, in which a data record (141)
is inserted.
The latter data record is taken from the ACCOUNT table (data record 65 of Fig.
3) and it has
a key attribute - Account No ="133333" (constituting the search key). This
data record,
despite taken from a different entity, is inserted to the PAIF 160 by strictly
following the above
referred to insert procedure.
Preferably, in order to better distinguish between data records of different
types that
reside in the same PAIF each data record belonging to a given type can be
prefixed with a
unique symbol. Thus, for example, all the search keys of data records that
belong to the
entity "CLIENT" are prefixed with the symbol 'A' (standing for "data type
ID"), whereas all the
search keys of data records that belong to the entity "ACCOUNT" are prefixed
with the
symbol 'B'. The new search key of the data records that belong to CLIENT
consists now of
the concatenation of 'A' and the original search key, and by the same token,
the new search
key of the data records that belong to ACCOUNT consists now of the
concatenation of'B' and
the original search key.
DOCSMTL: 2234878U
CA 02262593 2006-11-20
-34-
By following the latter approach all data records that belong to a given data
type
reside in a well defined "cluster" or sub PAIF within the PAIF The resulting
partitioning is
illustrated schematically in a PAIF 170 shown in Fig. 9. Thus, the symbol 'A'
and 'B' at offset 0
of the search key, inherently divide the data records into two distinct
clusters, respectively,
such that all data records belonging to the CLIENT entity reside in cluster
172 of the PAIF
and all data records that belong to "ACCOUNT" reside in cluster 174 of the
PAIF.
Thus, whilst according to hitherto known solutions, data of different types
are typically
held in different files, according to a database management system utilizing a
PAIF, different
data types may reside in the same file. Utilizing only one file is, typically
(but not necessarily),
advantageous over utilizing a plurality of files since it obviates the need to
repeatedly invoking
the relatively time consuming operating system services "open file" and "close
file", and
thereby improves the overall database response time. It also simplifies the
memory
management requirements, e.g. it enables the use a single cache memory for
blocks of the
entire system.
It should be noted that the search keys of data records that belong to
different types
(and reside in the same PAIF) do not necessarily have the same length. Thus,
in the example
of Fig. 9 all records that belong to the "CLIENT" entity have 5 digits,
whereas the record 141
that belong to the "ACCOUNT" entity has 6 digit. This characteristic
facilitates the application
of the file management system of the invention to databases indicative of wide
range of real-
life scenarios.
In a similar manner, data records that belong to the relationship "DEPOSIT"
may be
also incorporated in the PAIF 160 of Fig. 8, e.g. data record 70 from the
"DEPOSIT" table of
Fig. 3. The pertinent search key consists of the concatenation of the search
keys of "CLIENT"
and "ACCOUNT", i.e. "11346133333" (being indicative of the fact that "CLIENT"
11346 has
an account no. 133333). In order to distinguish data records that belong to
"DEPOSIT" from
those that belong to "ACCOUNT" and "CLIENT", the above tagging technique may
be
utilized, i.e. all data records that belong to the "DEPOSIT" table arc
prefixed with the symbol
'C' (as opposed to'A' and 'B' used for "CLIENT" and "ACCOUNT", respectively).
Accordingly,
the specified search key is "C11346133333". As is well known to those versed
in the art, for
the purpose of realizing efficient data manipulation it may be required, in
certain scenarios, to
view the specified record not as " a client having an account" but rather as
"an account
DOCSMTL: 2234878\1
CA 02262593 2006-11-20
- 35 -
owned by a client". To this end the search-key of data records belonging to
"DEPOSIT" may
also be viewed as a concatenation of "ACCOUNT" and "CLIENT" (e.g. the search
key of the
same data record 70 is also "13333311346"). In order to distinguish this
record from its
counterpart in the relationship "DEPOSIT" it is prefixed with the symbol 'D',
i.e.
"D 13333311346".
Storing both 011346133333 and D13333311346 in an PAIF is, obviously, undesired
since they represent the same data occurrence (data record 70). One possible,
yet not
exclusive approach to cope with this limitation is to construct a PAIF of the
kind shown in Fig.
10. Consider, for example, a PAIF 180 having a leaf node 182 pointing, by
means of long link
184, to data link 186 having a search key 'Axxxx' (from the "Account" table)
which is
composed of a concatenation of the letter 'A' and the contents of the field
"Account No."
Suppose that an account number has additional attributes e.g date (at which
the account was
opened) and others. The additional attributes are organized in a separate data
record (188)
having a search key that is composed of the concatenation of the letter'B' and
the contents of
the field "date" ("yyyy"). As shown in Fig. 10, according to an alternative
approach data record
188 is linked by means of a long link 190 and data record 186 is accessible by
means of
pointer 192. Put differently, the leaf node 182 is linked to link list which
consists of records
188 and 186. It is important to note that data record 188, albeit having a
search key "Byyyy"
is connected to a node that forms part of a search path that is defined by the
search key of
data record 186, i.e. 'Axxxx'. This structure whilst seemingly not complying
with the general
definition of PAIF is nevertheless appropriate seeing that record 188 is
meaningful only in
conjunction with record 186 (in a sense it is subordinated thereto). More
specifically the field
"opening date" in record 188 is meaningless per se, however it gains meaning
when viewed
as the "opening date" of the account having "account number" as defined in
record 186.
Accordingly one could possibly refer to record 188 as having a search key
composed of the
concatenation of the search keys of records 186 and 188, i.e. "AxxxByyyy".
Data record 186,
whilst not directly linked to node182, is associated thereto. The example of
Fig. 10, illustrates
a situation in which more than one data record is associated to a single node
in a PAIF.
Turning now to Fig. 11, there is shown a schematic illustration of a single
PAIF,
according to one embodiment of the invention, utilized for representing the
relationship
"DEPOSIT". Reverting to the above example the specific Account data record has
a search
DOCSMTL: 2234878\1
CA 02262593 2006-11-20
-36-
key 'A133333' (201 ) the specific Client has a data record 'B11346'(202). The
fact that a given
account number ("133333") is associated to a given Client ("11346") is
represented as
"C11346"(3,03). The latter data record is subordinated to data record 201 and
therefore it is
identified also by the search key of record 203. Put differently, one can
refer to the search key
of data record 203 as a concatenation of the search keys of data records201
and 203, i.e.
'A133333C 11346'.
By the same token, data record 2O4 represents an account of Client data record
202.
The search path that leads to data record 204 is identified by concatenation
of the search
keys of data record 202, i.e. 'B11346' and that of data record 204, i.e.
'D133333'.
The resulting PAIF 200 includes data record 201 and 203 where the latter is
linked by
means of long pointer 206 to a node 207. As specified before, the search path
that leads to
node 207 corresponds to the search key of data record 201.
The resulting PAIF (301) further includes data record 203, and 204 where the
latter is
linked by means of long pointer 208 to a node 209. As specified before, the
search path that
leads to node 209 corresponds to the concatenation of the search keys of data
records 202
and 204. As shown in PAIF200, the Client and Account data records are
duplicated which is
an obvious drawback which results in an undue inflated file.
This drawback may be easily overcome by representing the DEPOSIT relationship
as
structure 210. Thus, data records 201 and 203 that are linked to node 207 by
means of
pointer 206 and data records 202 and 204 that are linked to node 209 by means
of pointer
208, are represented as a single data structure 210.
As shown, the search path defined by the search keys of records 201 and 203
leads
to the first field212. having a value 'C'. The third field points to the
actual data record 201. The
second field 215 (having a value 'D') of the same data structure210 is
accessible by search
path that is defined by the search key of record 202. The fourth field has a
pointer that points
to the actual data record202. In this manner the relationship DEPOSIT is
represented both as
"a client having an account" and as an "account owned by a client", whilst
avoiding
duplication of the records, i.e. only one instance of the data records
"client" and "account" is
DOCSMTL: 2234878\1
CA 02262593 2006-11-20
-37-
used for representing the relationship DEPOSIT. Field 216 illustrates an
attribute of the
DEPOSIT relationship, e.g. the balance of the account "133333" owned by Client
"11346".
DOCSMTL: 2234878U
CA 02262593 1999-O1-26
WO 98/04980 PCT/8.97/00248
-38-
The PAIF is associated with additional table (not shown) that
describes the relations, i.e. that 'C' (field ? 12) is indicative of an
account
record that is physically linked to field 213 and that the client that owns
this
record is physically linked to field 214. 'D' (field 21~), in its turn, is
indicative of a client record that is physically linked to field ?14 and that
client has an account record which is physically linked to field 213.
Accordingly, there is only one occurrence of "Client" and "Account" in the
PAIF. The search path that leads to field 212 ('C') corresponds, of course,
to AxxxxByyyy, whereas the search path that leads to field ?1~ ('D')
corresponds to ByyyyAxxxx. If the sought data record is Axxxx (i.e. the
client record 201 per se), then one simply moves in the PAIF with a search
key of 'Axxxx', in the manner specified above, and reaches a given node
(e.g. node ?06'). From here one could continue to any leaf subordinated
thereto (by this particular example there is a sole subordinated leaf
1~ node (207)) and therefrom extract record 201 in the manner specified above,
or if desired there may be a direct link from node ?06' to data record ?O1.
Other implementation are of course feasible, all as required and appropriate.
In a PAIF 200 according to this embodiment, a leaf node is
associated with more than one data record (which by this particular
embodiment is represented as a sole data structure ? 10). As shown, the
PAIF 200 includes one physical occurrence of two data records (201
and 202), such that there are defined two search paths that lead to the same
two data records.
It should be noted that the PAIF of Figs. 10 and 11 illustrate
only one out of many possible variants of representing efficiently a
relationship in a PAIF.
Preferably, the database file management system of the
invention should be associated with known per se concurrency and/or
distributed capabilities so as to enable a plurality of users to access
virtually
simultaneously to the database. The database may be located in a central
location, or distributed among two or more remote locations.
CA 02262593 1999-O1-26
WO 98104980 PCT/8.97100248
-39-
Turning now to Figs. 1?A-D, there are shown four benchmark
graphs demonstrating the enhanced performance, in terms of response time
and file size of database utilizing a file management system of the invention
vs. commercially available Ctree based database. The inserts are realized
through Uniface application running in Windows (for workgroup) operating
system.
The benchmark of Fig. 1?A concerns measuring the time in
minutes for inserting an ever increasing number of a priori sorted data
records to a file (0-1,000,000). As shown in Fig. 1?A, the larger number of
inserts the greater is the improvement in terms of response time of the
database system of the invention. Thus inserting 1 million records takes
about 669 minutes in the Ctrec based database as compared to only 6~
minutes in the system of the invention. Moreover, the response time in the
file management system of the invention increases by only a small extent as
the number of records increases, as opposed to significant increase in the
response time in the counterpart system according to the prior art.
The benchmark of Fig. 12B illustrates the file size in mega
bytes as a function of number of data records in the file (0-1,000,000). As
shown in Fig. I?B, the larger number of records the greater is the improve-
?0 ment in terms of file size in the database system of the invention. Thus
for 1 million records the file size of Ctree based file is about 1~1 mega byte
as compared to only ?? mega byte in the database of the invention.
Graphs 12C and 1?D are similar to those shown in Figs. 12A
and 12B apart from the fact that in the former (1''C and 1?D) the data
2~ records are inserted randomally whereas in the latter (I?A and 12B) the
data
records are a priori sorted according to search key. As shown the results are
as before i.e. the system of the invention is more efficient in terms of both
response time and file size.
Figs. 13A-D illustrates a benchmark graphs of a system of the
30 invention (operating under DOS operating system) vs. commercially
available Btree based database system. The results are as before i.e. the
CA 02262593 1999-O1-26
WO 98/04980 PCT/8,97/00248
-4(~-
system of the invention is more efficient in terms of both response time and
file size.
The present invention has been described with a certain degree
of particularity but it should be understood that various modifications and
alteration may be made without departing from the scope or spirit of the
invention as defined by the following claims:
