Re^2: Distiguishing arguments: number-strings vs real integer

[LanX]

I'm hacking a DBI abstraction, and I'm not sure about the side-effects of binding 42 vs "42" to a placeholder.

So I want to play safe and keep full control.

see also should-i-quote-numbers-in-sql

I hope that DBI handles that automatically according to the type of the column, but again I'm not sure ...

[choroba]

It depends on the driver, see Why does DBI implicitly change integers to strings?

[bliako]

I'm hacking a DBI abstraction, and I'm not sure about the side-effects of binding 42 vs "42" to a placeholder.

Isn't irrelevant what Perl thinks of a variable's type when DB already knows/was told what the type of that column should be at create time? If DB column was declared as INT then don't quote (and possibly warn user of a type mismatch if any), else quote.

Having said that I read in Programming the Perl DBI:

It's equally simple to specify multiple bind values within one statement, since bind_ param( ) takes the index, starting from 1, of the parameter to bind the given value to. For example:

    $sth = $dbh->prepare( "
                SELECT name, location
                FROM megaliths
                WHERE name = ?
                AND mapref = ?
                AND type LIKE ?
            " );
    $sth->bind_param( 1, "Avebury" );
    $sth->bind_param( 2, $mapreference );
    $sth->bind_param( 3, "%Stone Circle%" );

You may have noticed that we haven't called the quote( ) method on the values. Bind values are passed to the database separately from the SQL statement,[50] so there's no need to "wrap up" the value in SQL quoting rules.

    [50] This is not strictly true, since some drivers emulate placeholders by doing a textual replacement of the placeholders with bind values before passing the SQL to the database. Such drivers use Perl's internal information to guess whether each value needs quoting or not. Refer to the driver documentation for more information. [emphasis mine]

[LanX]

> Isn't irrelevant what Perl thinks of a variable's type when DB already knows/was told what the type of that column should be at create time?

some testing on MariaDB

CREATE TABLE `t_test_type` (
    `f_num` INT(11) NULL DEFAULT NULL,
    `f_str` CHAR(50) NULL DEFAULT NULL
)
ENGINE=MyISAM
;

select * from t_test_type;
1;1
2;2
3;3

select * from t_test_type where f_str < "2.0";
1;1
2;2


select * from t_test_type where f_str < 2.0;
1;1

[mr_ron]

> Isn't irrelevant what Perl thinks of a variable's type when DB already knows/was told what the type of that column should be at create time? some testing on MariaDB

CREATE TABLE `t_test_type` (
    `f_num` INT(11) NULL DEFAULT NULL,
    `f_str` CHAR(50) NULL DEFAULT NULL
)
ENGINE=MyISAM
;

It looks like you could get what you want from the ansi information_schema and a query like the one below that worked for me under MySQL. I have read that not all DBMS support information_schema.

> SELECT column_name, data_type FROM information_schema.columns
WHERE table_name = 't_test_type' AND column_name IN ('f_num', 'f_str')
+;

+-------------+-----------+
| column_name | data_type |
+-------------+-----------+
| f_num       | int       |
| f_str       | char      |
+-------------+-----------+
2 rows in set (0.00 sec)

DBI has a column_info method that seems to guarantee DATA_TYPE code and TYPE_NAME fields, if implemented by the driver which should be true for MariaDB based (only) on looking at the source.

#!/usr/bin/env perl

use warnings;
use strict;

use DBI;

my $dsn = "DBI:mysql:database=mysql";
my $dbh = DBI->connect($dsn,
    ... , #user name and password
    { RaiseError => 1 }
) or die $DBI::errstr;

my $info = $dbh->column_info(
    undef, 'mysql', 't_test_type', '%'
)->fetchall_hashref('COLUMN_NAME');
while (my ($col, $col_info) = each %$info) {
    print "$col $col_info->{ DATA_TYPE } $col_info->{ TYPE_NAME }\n";
}

Output:

f_str 1 CHAR
f_num 4 INT

[syphilis]

I'm hacking a DBI abstraction, and I'm not sure about the side-effects of binding 42 vs "42" to a placeholder

Therefore, you need to determine whether the placeholder needs to be 42 or "42".

So I want to play safe and keep full control

Therefore, either:
1) having determined that the placeholder needs to be an integer, you provide $x + 0;
or
2) having determined that the placeholder needs to be a string, you provide "$x".

But I don't see that you would need to determine whether $x is an integer or a string.
Of course, things get a little more complex if there's a need to check the validity of the integer value, but that doesn't appear to be an issue (going by your description of the problem).

Cheers,
Rob

[LanX]

I find SQL very confusing and archaic that's why I want a safer abstraction layer.

Saying this it becomes even more confusing because of different dialects and differently implemented DB drivers.

> Therefore, you need to determine whether the placeholder needs to be 42 or "42".

This is up to the user of the layer, that's why I need to know if he passed a literal string or number. And to be able to warn him if necessary.

[syphilis]

This is up to the user of the layer, that's why I need to know if he passed a literal string or number. And to be able to warn him if necessary.

Ok ... so you firstly need to know which option the user has chosen.
Having determined that, I would have thought you could just pass on that user-supplied argument as "$x" (if string was chosen) or "$x"+0 (if integer was chosen) without any need to analyse just what that argument was.

But I've never worked with databases so you should probably simply ignore me (and perhaps even be envious ;-)

Cheers,
Rob