DBD::Sprite - Perl extension for DBI, providing database emmulation via flat files.  


    This module is Copyright (C) 2000, 2001, 2002, 2003, 2004 by

                Jim Turner

    All rights reserved.

    You may distribute this module under the terms of either the GNU General
    Public License or the Artistic License, as specified in the Perl README

        DBD::Sprite is a derived work by Jim Turner from, a module 
        written and copyrighted (c) 1995-1998, by Shishir Gurdavaram 


     use DBI;
     $dbh = DBI->connect("DBI:Sprite:spritedb",'user','password')
         or die "Cannot connect: " . $DBI::errstr;
     $sth = $dbh->prepare("CREATE TABLE a (id INTEGER, name CHAR(10))")
         or die "Cannot prepare: " . $dbh->errstr();
     $sth->execute() or die "Cannot execute: " . $sth->errstr();


DBD::Sprite is a DBI extension module adding database emulation via flat-files or XML documents to Perl's database-independent database interface. Unlike other DBD::modules, DBD::Sprite does not require you to purchase or obtain a database. Every thing you need to prototype database-independent applications using Perl and DBI are included here. You will, however, probably wish to obtain a real database, such as "mysql", for your production and larger data needs. This is because emulating databases and SQL with flat text files gets very slow as the size of your "database" grows to a non-trivial size (a few thousand records or so per table).

DBD::Sprite is built upon an old Perl module called "Sprite", written by Shishir Gurdavaram. This code was used as a starting point. It was completly reworked and many new features were added, producing a module called "" (Jim Turner's Sprite). This was then merged in to DBI::DBD to produce what you are installing now. (DBD::Sprite). is included in this module as a separate file, and is required.

Many thanks go to Mr. Gurdavaram.

The main advantage of DBD::Sprite is the ability to develop and test prototype applications on personal machines (or other machines which do not have an Oracle licence or some other "mainstream" database) before releasing them on "production" machines which do have a "real" database. This can all be done with minimal or no changes to your Perl code.

Another advantage of DBD::Sprite is that you can use Perl's regular expressions in your SQL "WHERE-clauses" to search through your data (ie. "...where FIELD =~ /\S/..."). Maybe, someday, more "real" databases will include this feature too!

DBD::Sprite provides the ability to emulate basic database tables and SQL calls via flat-files. The primary use envisioned for this to permit website developers who can not afford to purchase an Oracle licence to prototype and develop Perl applications on their own equipment for later hosting at larger customer sites where Oracle is used. :-)

DBD::Sprite attempts to do things in as database-independent manner as possible, but where differences occurr, DBD::Sprite most closely emmulates Oracle, for example "sequences/autonumbering. DBD::Sprite uses tiny one-line text files called `"sequence files (.seq). and `"seq_file_name.NEXTVAL function to insert into autonumbered fields. The reason for this is that the Author works in an Oracle shop and wrote this module to allow himself to work on code on his PC, and machines which did not have Oracle on them, since obtaining Oracle licences was sometimes time-consuming.

DBD::Sprite is similar to DBD::CSV, but differs in the following ways:


    Installing this module (and the prerequisites from above) is quite
    simple. You just fetch the archive, extract it with

        gzip -cd DBD-Sprite-0.1000.tar.gz | tar xf -

    (this is for Unix users, Windows users would prefer WinZip or something
    similar) and then enter the following:

        cd DBD-Sprite-#.###
        perl Makefile.PL
        make test

    If any tests fail, let me know. Otherwise go on with

        make install

    Note that you almost definitely need root or administrator permissions.
    If you don't have them, read the ExtUtils::MakeMaker man page for
    details on installing in your own directories. the ExtUtils::MakeMaker

        NOTE:  You will also need to copy "" to /usr/local/bin or 
        somewhere in your path.

Windows installation

		If installing in Windows, you must 1st install the DBI module, 
	create a DBD subdirectory in your Perl's path (run "perl -V" to find out 
	what this is), copy "" to it, then copy the other files (,, and to the same directory you 
	created the DBD subdirectory in.  Then copy the file to the 
	directory perl itself is in.  These directories (in ActivePerl) are: 
	c:\perl\site\lib and c:\perl\bin respectively.  You do NOT need to use 
	"make" or "nmake" to build DBD::Sprite!


        1) cd to where you wish to store your database.
        2) run to create your database, ie.
                Database name: mydb
                Database user: me
                User password: mypassword
                Database path: .
                Table file extension (default .stb): 
                Record delimiter (default \n): 
                Field delimiter (default ::): 

                This will create a new database text file (mydb.sdb) in the current 
                directory.  This ascii file contains the information you enterred 
                above.  To add additional user-spaces, simply rerun with 
                "mydb" as your database name, and enter additional users (name, 
                password, path, extension, and delimiters).  For an example, after 
                running "make test", look at the file "test.sdb".               
                When connecting to a Sprite database, Sprite will look in the current 
                directory, then, if specified, the path in the SPRITE_HOME environment 

                The database name, user, and password are used in the "db->connect()" 
                method described below.  The "database path" is where your tables will 
                be created and reside.  Table files are ascii text files which will 
                have, by default, the extension ".stb" (Sprite table).  By default, 
                each record will be written to a single line (separated by \n -- 
                Windows users should probably use "\r\n").  Each field datum will be 
                written without quotes separated by the "field delimiter (default: 
                double-colon).  The first line of the table file consists of the 
                a field name, an equal ("=") sign, an asterisk if it is a key field, 
                then the datatype and size.  This information is included for each 
                field and separated by the field separator.  For an example, after 
                running "make test", look at the file "testtable.stb".          

        3) write your script to use DBI, ie:
                use DBI;
                $dbh = DBI->connect('DBI:Sprite:mydb','me','mypassword') || 
                                die "Could not connect (".$DBI->err.':'.$DBI->errstr.")!";
        4) get your application working.
        5) rehost your application on a "production" machine and change "Sprite" 
        to a DBI driver for a "real" database!


    You can create and drop tables with commands like the following:

        $dbh->do("CREATE TABLE $table (id INTEGER, name CHAR(64))");
        $dbh->do("DROP TABLE $table");

    A drop just removes the file without any warning.

    See the DBI(3) manpage for more details.

    Table names cannot be arbitrary, due to restrictions of the SQL syntax.
    I recommend that table names are valid SQL identifiers: The first
    character is alphabetic, followed by an arbitrary number of alphanumeric
    characters. If you want to use other files, the file names must start
    with '/', './' or '../' and they must not contain white space.


    The following examples insert some data in a table and fetch it back:
    First all data in the string:

        $dbh->do("INSERT INTO $table VALUES (1, 'foobar')");

    Note the use of the quote method for escaping the word 'foobar'. Any
    string must be escaped, even if it doesn't contain binary data.

    Next an example using parameters:

        $dbh->do("INSERT INTO $table VALUES (?, ?)", undef,
                 2, "It's a string!");

    To retrieve data, you can use the following:

        my($query) = "SELECT * FROM $table WHERE id > 1 ORDER BY id";
        my($sth) = $dbh->prepare($query);
        while (my $row = $sth->fetchrow_hashref) {
            print("Found result row: id = ", $row->{'id'},
                  ", name = ", $row->{'name'});

    Again, column binding works: The same example again.

        my($query) = "SELECT * FROM $table WHERE id > 1 ORDER BY id";
        my($sth) = $dbh->prepare($query);
        my($id, $name);
        $sth->bind_columns(undef, \$id, \$name);
        while ($sth->fetch) {
            print("Found result row: id = $id, name = $name\n");

    Of course you can even use input parameters. Here's the same example for
    the third time:

        my($query) = "SELECT * FROM $table WHERE id = ?";
        my($sth) = $dbh->prepare($query);
        $sth->bind_columns(undef, \$id, \$name);
        for (my($i) = 1;  $i <= 2;   $i++) {
            if ($sth->fetch) {
                print("Found result row: id = $id, name = $name\n");

    See the DBI(3) manpage for details on these methods. See the
    SQL::Statement(3) manpage for details on the WHERE clause.

    Data rows are modified with the UPDATE statement:

        $dbh->do("UPDATE $table SET id = 3 WHERE id = 1");

    Likewise you use the DELETE statement for removing rows:

        $dbh->do("DELETE FROM $table WHERE id > 1");


Method takes 2 arguments: Function name and optionally, a package name (default is ""main").

                $dbh->fn_register ('myfn','mypackage');

                use JSprite;
                JSprite::fn_register ('myfn',__PACKAGE__);

Then, you could say in sql:

        insert into mytable values (myfn(?))
and bind some value to "?", which is passed to "myfn", and the return-value 
is inserted into the database.  You could also say (without binding):

        insert into mytable values (myfn('mystring'))
-or (if the function takes a number)-

        select field1, field2 from mytable where field3 = myfn(123) 
I<Return Value>



As of v. 0.50, basic two-table inner-joins are now supported. For example:

        select t1.field1, t2.field1, t1.field2 from table1 t1, table2 t2
        where t1.field1 = t2.field3 order by t2.field1 desc, t1.field1
This would return the three fields requested based on a set intersection of all records in table1 and table2 such that field1 of table1 matches field3 of table3. NOTE: This is the ONLY type of join currently supported! You can, however add additional selection criteria and or ordering arguments. You can also omit the where-clause and get a "set union" of the specified fields for all records of both tables.


    In the above examples we have never cared about return codes. Of course,
    this cannot be recommended. Instead we should have written (for

        my($query) = "SELECT * FROM $table WHERE id = ?";
        my($sth) = $dbh->prepare($query)
            or die "prepare: " . $dbh->errstr();
        $sth->bind_columns(undef, \$id, \$name)
            or die "bind_columns: " . $dbh->errstr();
        for (my($i) = 1;  $i <= 2;   $i++) {
                or die "execute: " . $dbh->errstr();
            if ($sth->fetch) {
                print("Found result row: id = $id, name = $name\n");
            or die "finish: " . $dbh->errstr();

    Obviously this is tedious. Fortunately we have DBI's *RaiseError*

        $dbh->{'RaiseError'} = 1;
        $@ = '';
        eval {
            my($query) = "SELECT * FROM $table WHERE id = ?";
            my($sth) = $dbh->prepare($query);
            $sth->bind_columns(undef, \$id, \$name);
            for (my($i) = 1;  $i <= 2;   $i++) {
                if ($sth->fetch) {
                    print("Found result row: id = $id, name = $name\n");
        if ($@) { die "SQL database error: $@"; }

    This is not only shorter, it even works when using DBI methods within


    The following attributes are handled by DBI itself and not by DBD::Sprite,
    thus they should all work as expected:  I have only used the last 3.

        CompatMode             (Not used)

    The following DBI attributes are handled by DBD::Sprite:


        Should Work

        Valid after `$sth->execute'

        Valid after `$sth->prepare'

        Valid after `$sth->execute'; undef for Non-Select statements.

        Not really working. Always returns a reference to an array of '1's, as
        DBD::Sprite always allows NULL (handled as an empty string). 
        Valid after `$sth->execute'.

                Should work

		    		works, except setting to zero allows any length of data to be 

    These attributes and methods are not supported:


    In addition to the DBI attributes, you can use the following dbh
    attributes.  These attributes are read-only after "connect".

    The following are Sprite-specific options which can be set when connecting.

            Path to tables for database.
        File extension used on table files in the database.
        Current database user.

        Field delimiter string in use for the database.
        Default specified in database configuration file (.sdb)
        Field delimiter string in use for inputting the database.
        Default = sprite_field
        Field delimiter string in use for outputting the database.
        Default = sprite_field
    sprite_xsl  (NEW)
        Allows specifying of a url to an xsl template to be written to xml 
        documents (when using the "xml" option).  This makes it very easy to 
        view your tables via M$ Internet Explorer browser!

        Example:  sprite_xsl => ''
        Default is none.  Only applies if "sprite_field" is set to "xml"!

    sprite_dbfdelim - DEPRECIATED, now use "sprite_field"!
        Field delimiter string in use for the database.
    sprite_dbrdelim - DEPRECIATED, now use "sprite_record"!
        Record delimiter string in use for the database.
        By default, table names are case-insensitive (as they are in Oracle),
        to make table names case-sensitive (as in MySql), so that one could
        have two separate tables such as "test" and "TEST", set this option
        to 1.

    sprite_CaseFieldNames  (NEW)
        By default, field names are case-insensitive (as they are in Oracle),
        to make field names case-sensitive (as in XML), so that one could
        have two separate fields such as "field1" and "Field1", set this option
        to 1.

        "0" by defalt.  Specifies that encryption is to be used when storing 
        the data in the flat-file.  To use, download "Crypt::CBC", and one 
        or more of "Crypt::DES", "Crypt::IDEA", or "Crypt::Blowfish".  You 
        can specify using any of the following formats:
            sprite_Crypt => 'my key string'
                Use Blowfish encryption.
            sprite_Crypt => 'DES;my key string'
                Use DES encryption.
            sprite_Crypt => 'encrypt=CBC;IDEA;my key string'
                Use IDEA encription, but read in table as unencrypted, then 
                write it out encrypted (great for encrypting previously 
                unencrypted tables).
            sprite_Crypt => 'decrypt=CBC;Blowfish;my key string'
                use Blowfish encryption, but write out table unencrypted. 
                This allows one to fetch an encrypted table and write it back 
                out unencrypted.

    sprite_reclimit  (aka. sprite_sizelimit)
        Allows user to specify the maximum number of records to be returned 
        by a single query.  Default is "0", which permits an unlimited number.

        CHAR fields are always right-padded with spaces to fill out
        the field.  Old (pre 5.17) Sprite behaviour was to require the
        padding be included in literals used for testing equality in
        "where" clauses.    I discovered that Oracle and some other databases
        do not require this when testing DBIx-Recordset, so Sprite will
        automatically right-pad literals when testing for equality.
        To disable this and force the old behavior, set this option to 1.

        Specify alternate path/file to use for file-locking on os'es without 
        a working "flock" (ie. M$ os'es).  Default is 
        "/path/to/where/my/tables/are/Sprite.lck".  This is useful if the 
        file-system your tables are on is read-only.

        Allows user-specification of the number of times to try to create 
        lock-file on os'es without a working "flock" (ie. M$ os'es).  Default 
        is 10.

        Forces existing Sprite table file to be deleted and recreated in lieu 
        of being overwritten.  This was necessary because the author uses a 
        Samba filesystem which for some reason does not permit overwriting 
        of files.

        Environment variable specifying a path to search for Sprite 
        databases (*.sdb) files.


        The `data_sources' method returns a list of "databases" (.sdb files) 
        found in the current directory and, if specified, the path in 
        the SPRITE_HOME environment variable.
        This method returns a list of table names specified in the current 

            my($dbh) = DBI->connect("DBI:Sprite:mydatabase",'me','mypswd');
            my(@list) = $dbh->func('tables');

    JSprite::fn_register ('myfn',__PACKAGE__);
        This method takes the name of a user-defined data-conversion function 
        for use in SQL commands.  Your function can optionally take arguments, 
        but should return a single number or string.  Unless your function 
        is defined in package "main", you must also specify the package name 
        or "__PACKAGE__" for the current package.  For an example, see the 
        section "INSERTING, FETCHING AND MODIFYING DATA" above or (JSprite(3)).

                This utility lets you build new Sprite databases and later add 
                additional user-spaces to them.  Simply cd to the directory where 
                you wish to create / modify a database, and run.  It prompts as 
                Database name: Enter a 1-word name for your database.
                Database user: Enter a 1-word user-name.
                User password: Enter a 1-word password for this user.
                Database path: Enter a path (no trailing backslash) to store tables.
                Table file extension (default .stb): 
                Record delimiter (default \n): 
                Field delimiter (default ::): 

                The last 6 prompts repeat until you do not enter another user-name 
                allowing you to set up multiple users in a single database.  Each 
                "user" can have it's own separate tables by specifying different 
                paths, file-extensions, password, and delimiters!  You can invoke 
                "" on an existing database to add new users.  You can 
                edit it with vi to remove users, delete the 5 lines starting with 
                the path for that user.  The file is all text, except for the 
                password, which is encrypted for your protection!


        Although DBD::Sprite supports the following datatypes:
                DATE LONG BLOB MEMO and RAW, there are really only 4 basic datatypes 
                (NUMBER, CHAR, VARCHAR, and BLOB).  This is because Perl treates 
                everything as simple strings.  The first 6 are all treated as "numbers" 
                by Perl for sorting purposes and the rest as strings.  This is seen 
                when sorting, ie NUMERIC types sort as 1,5,10,40,200, whereas 
                STRING types sort these as 1,10,200,40,5.  CHAR fields are right-
                padded with spaces when stored.  LONG-type fields are subject to 
                truncation by the "LongReadLen" attribute value.  BLOB-type 
                fields have their data stored on separate files created by Sprite.

        DBD::Sprite works with the tieDBI module, if "Sprite => 1" lines are added 
        to the "%CAN_BIND" and "%CAN_BINDSELECT" hashes.  This should not be 
        necessary, and I will investigate when I have time.

	Data field names must NOT be the same as any of Perl's string operators 
	(eq lt le ge gt, etc) or boolean operators (and or not)!



Revision history for Perl extension DBD::Sprite.


Extensions of DBD::Sprite

    *       Additional Oracle-ish functions built-in as requested.

    *       Delete dangling blob-files when dropping a table.


    *       The module is using flock() internally. However, this function is
            not available on platforms. Using flock() is disabled on MacOS
            and Windows 95: There's no locking at all (perhaps not so
            important on these operating systems, as they are for single
            users anyways).

    *       Unique-key violations on updates probably will not be caught 
            if the argument is a function-call or a field-name.

    *       Sorting for joins:  In order to obtain correct sorting, all 
            fields specifying desired sort order must be grouped together by 
            table, ie. "order by t1.field1, t1.field2, t2.field1, t2.field2", 
            which should work fine.  This is due to the fact that the records 
            are merged together after data for each table is sorted 
            independently.  This should not be an issue for 95+% of all sorts,  
            If it is, patches are welcome and my very own "sort_elements()" 
            module is available to you!  An example of a sort known NOT to 
            work exactly right would be:  
            "order by t1.field1, t2.field1, t1.field2".


        JSprite(3), DBI(3), perl(1)