SQL::Eval - Base for deriving evaluation objects for SQL::Statement
require SQL::Statement;
require SQL::Eval;
# Create an SQL statement; use a concrete subclass of
# SQL::Statement
my $stmt = MyStatement->new("SELECT * FROM foo, bar",
SQL::Parser->new('Ansi'));
# Get an eval object by calling open_tables; this
# will call MyStatement::open_table
my $eval = $stmt->open_tables($data);
# Set parameter 0 to 'Van Gogh'
$eval->param(0, 'Van Gogh');
# Get parameter 2
my $param = $eval->param(2);
# Get the SQL::Eval::Table object referring the 'foo' table
my $fooTable = $eval->table('foo');
This module implements two classes that can be used for deriving
subclasses to evaluate SQL::Statement objects. The SQL::Eval object
can be thought as an abstract state engine for executing SQL queries
and the SQL::Eval::Table object is a table abstraction. It implements
methods for fetching or storing rows, retrieving column names and
numbers and so on. See the test.pl script as an example for
implementing a subclass.
While reading on, keep in mind that these are abstract classes,
you *must* implement at least some of the methods described below.
In addition, you need not derive from SQL::Eval or SQL::Eval::Table,
you just need to implement the method interface.
All methods throw a Perl exception in case of errors.
- new
-
Constructor; use it like this:
$eval = SQL::Eval->new(\%attr);
Blesses the hash ref \%attr into the SQL::Eval class (or a subclass).
- param
-
Used for getting or setting input parameters, as in the SQL query
INSERT INTO foo VALUES (?, ?);
Example:
$eval->param(0, $val); # Set parameter 0
$eval->param(0); # Get parameter 0
- params
-
Used for getting or setting the complete array of input
parameters. Example:
$eval->params($params); # Set the array
$eval->params(); # Get the array
- table
-
Returns or sets a table object. Example:
$eval->table('foo', $fooTable); # Set the 'foo' table object
$eval->table('foo'); # Return the 'foo' table object
- column
-
Return the value of a column with a given name; example:
$col = $eval->column('foo', 'id'); # Return the 'id' column of
# the current row in the
# 'foo' table
This is equivalent to and a shorthand for
$col = $eval->table('foo')->column('id');
- _gen_access_fastpath
-
Return a subroutine reference for fast accessing columns for read-only
access. This routine simply returns the
_gen_access_fastpath of the
referenced table.
- newnew
-
Constructor; use it like this:
$eval = SQL::Eval::Table->new(\%attr);
Blesses the hash ref \%attr into the SQL::Eval::Table class (or a
subclass).
The following attributes are used by SQL::Eval::Table:
- col_names
-
Array reference containing the names of the columns in order they appear
in the table. This attribute must be provided by the derived class.
- col_nums
-
Hash reference containing the column names as keys and the column
indexes as values. If this is omitted (does not exist), it will be
created from
col_names.
- capabilities
-
Hash reference containing additional capabilities.
- _gen_access_fastpath_gen_access_fastpath
-
Return a subroutine reference for fast accessing columns for read-only
access. When the instantiated object does not provide own methods for
column and column_num a subroutine reference is returned which
directly access the internal data structures. For all other cases a
subroutine directly calling $self->column($_[0]) is returned.
- row
-
Used to get the current row as an array ref. Do not confuse
getting the current row with the fetch_row method! In fact this
method is valid only after a successful
$table->fetchrow().
Example:
$row = $table->row();
- columncolumn
-
Get the column with a given name in the current row. Valid only after
a successful
$table->fetchrow(). Example:
$col = $table->column($colName);
- column_num
-
Return the number of the given column name. Column numbers start with
0. Returns undef, if a column name is not defined, so that you can use
this for verifying column names. Example:
$colNum = $table->column_num($colNum);
- col_numscol_nums
-
Returns an hash ref of column names with the column names as keys and
the column indexes as the values.
- col_namescol_names
-
Returns an array ref of column names ordered by their index within the table.
- capability
-
Returns a boolean value whether the table has the specified capability
or not. This method might be overridden by derived classes, but ensure
that in that case the parent capability method is called when the
derived class does not handle the requested capability.
The following capabilities are used (and requested) by SQL::Statement:
- update_one_row
-
Defines whether the table is able to update one single row. This
capability is used for backward compatibility and might have
(depending on table implementation) several limitations. Please
carefully study the documentation of the table or ask the author of
the table, if this information is not provided.
This capability is evaluated automatically on first request and must
not be handled by any derived classes.
- update_specific_row
-
Defines if the table is able to update one single row, but keeps the
original content of the row to update.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
- update_current_row
-
Defines if the table is able to update the currently touched row. This
capability requires the capability of
inplace_update.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
- rowwise_update
-
Defines if the table is able to do row-wise updates which means one
of
update_one_row, update_specific_row or update_current_row.
The update_current_row is only evaluated if the table has the
inplace_update capability.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
- inplace_update
-
Defines if an update of a row has side effects (capability is not
available) or can be done without harming any other currently running
task on the table.
Example: The table storage is using a hash on the PRIMARY KEY of
the table. Real perl hashes do not care when an item is updated while
the hash is traversed using each. SDBM_File 1.06 has a bug,
which does not adjust the traversal pointer when an item is deleted.
SQL::Statement::RAM::Table recognizes such situations and adjusts
the traversal pointer.
This might not be possible for all implementations which can update
single rows.
This capability could be provided by a derived class only.
- delete_one_row
-
Defines whether the table can delete one single row by it's content or
not.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
- delete_current_row
-
Defines whether a table can delete the current traversed row or
not. This capability requires the
inplace_delete capability.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
- rowwise_delete
-
Defines if any row-wise delete operation is provided by the
table.
row-wise delete capabilities are delete_one_row and
delete_current_row.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
- inplace_delete
-
Defines if the deletion of a row has side effects (capability is not
available) or can be done without harming any other currently running
task on the table.
This capability should be provided by a derived class only.
- insert_new_row
-
Defines if a table can easily insert a new row without need to seek
or truncate. This capability is provided by defining the table class
method
insert_new_row.
This capability is evaluated automatically on first request and must not
be handled by derived classes.
If the capabilities rowwise_update and insert_new_row are
provided, the table primitive push_row is not required anymore and
may be omitted.
The above methods are implemented by SQL::Eval::Table. The following
methods are not, so that they *must* be implemented by the
subclass. See the DBD::DBM::Table or DBD::CSV::Table for
example.
- drop
-
Drops the table. All resources allocated by the table must be released
after
$table-drop($data)>.
- fetch_row
-
Fetches the next row from the table. Returns
undef, if the last
row was already fetched. The argument $data is for private use of
the subclass. Example:
$row = $table->fetch_row($data);
Note, that you may use
$row = $table->row();
for retrieving the same row again, until the next call of fetch_row.
SQL::Statement requires that the last fetched row is available again
and again via $table-row()>.
- push_row
-
As fetch_row except for storing rows. Example:
$table->push_row($data, $row);
- push_names
-
Used by the CREATE TABLE statement to set the column names of the
new table. Receives an array ref of names. Example:
$table->push_names($data, $names);
- seek
-
Similar to the seek method of a filehandle; used for setting the number
of the next row being written. Example:
$table->seek($data, $whence, $rowNum);
Actually the current implementation only uses seek($data, 0, 0)
(first row) and seek($data, 2, 0) (beyond last row, end of file).
- truncate
-
Truncates a table after the current row. Example:
$table->truncate($data);
The current implementation is quite simple: An SQL::Eval object is an
hash ref with only two attributes. The params attribute is an array
ref of parameters. The tables attribute is an hash ref of table
names (keys) and table objects (values).
SQL::Eval::Table instances are implemented as hash refs. Attributes
used are row (the array ref of the current row), col_nums (an
hash ref of column names as keys and column numbers as values) and
col_names, an array ref of column names with the column numbers as
indexes.
All methods are working with instance-local data only, thus the module
is reentrant and thread safe, if you either do not share handles between
threads or grant serialized use.
Please report any bugs or feature requests to bug-sql-statement at
rt.cpan.org, or through the web interface at
http://rt.cpan.org/NoAuth/ReportBug.html. I
will be notified, and then you will automatically be notified of
progress on your bug as I make changes.
You can find documentation for this module with the perldoc command.
perldoc SQL::Eval
perldoc SQL::Statement
You can also look for information at:
Written by Jochen Wiedmann and currently maintained by Jens Rehsack.
This module is Copyright (C) 1998 by
Jochen Wiedmann
Am Eisteich 9
72555 Metzingen
Germany
Email: joe@ispsoft.de
Phone: +49 7123 14887
and Copyright (C) 2009, 2017 by
Jens Rehsack < rehsackATcpan.org>
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 file.
the SQL::Statement(3) manpage
|