Description: |
I needed a way to get name/value pairs from a table using another column as the "object ID." This allowed me to have access to the values in a hash instead of having to do multiple selects.
Warning: this reads the entire data set into memory, so be careful not to read huge amounts of data.
I'm using &get_hash(...) to create a hash filled with values to fill a template created with HTML::Template, I'm also fetching the template from another table in my real script; I simplified the snippet by putting the template in with the attributes with a lookup of "page". With this method, I can use two select statements with one connect statement to create a dynamic web page instead of a file load and several select statements. :)
Update: added {RaiseError=>1}, put parameter binding in get_hash(...), and added array slice in $sth->fetchall_arrayref().
--
hiseldl |
#!/usr/bin/perl -w
use strict;
use DBI;
use HTML::Template;
####################################################
# Example usage
####################################################
# connect to the database
my $dbh = DBI->connect("dbi:mysql:mydb", "user", "pass",
{RaiseError=>1} );
# create the hash from the name/value pairs in the db
# using object_id="1"
my %params = &get_hash($dbh, "hashobjects", 1, 2, "object_id", "1" );
# create the html template with the page value from the DB
my $template = HTML::Template->new(scalarref => \$params{page},
die_on_bad_params => 0,
);
# substitute the params in the template
$template->param(\%params);
# print everything out
print "Content-Type: text/html\n\n";
print $template->output;
# disconnect so we don't get the warning,
# "Database handle destroyed without explicit disconnect."
$dbh->disconnect();
exit 0;
####################################################
# END Example usage
####################################################
####################################################
####################################################
# get_hash
# params:
# $dbhandle = database handle created with DBI->connect()
# $name_col = array index of column in the record set
# used for the key of the hash
# $val_col = array index of column in the record set
# used for the value of the hash
# $key_col = column name in the record set
# used for the key of the object id
# (this is the column name used in the WHERE clause)
# $key_val = column name in the record set
# used for the value of the object id
# (this is the column value used in the WHERE clause)
#
sub get_hash {
my ($dbhandle,$table,$name_col,$val_col,$key_col,$key_val)=@_;
my $stmt = "SELECT * FROM $table WHERE $key_col = ?";
my $sth = $dbhandle->prepare( $stmt );
$sth->execute($key_val);
# Here's where we read in the entire data set into
# an array ref containing refs to record arrays.
#my $ref=$sth->fetchall_arrayref();
my $ref = $sth->fetchall_arrayref([$name_col,$val_col]);
$sth->finish();
# use map to turn the array ref of array refs
# into a hash and then return it anonymously
#return map { $_->[$name_col]=>$_->[$val_col] } @$ref;
return map {@$_} @$ref;
}
####################################################
####################################################
__END__
# Here is a simplified version of the table
# along with some data to make the example
# run. This should work in MySQL.
# hashobjects.sql
#
# 1. save this data to a file
# 2. to create the table and view the data,
# at the mysql prompt type:
# mysql> use mydb;
# mysql> \. hashobjects.sql
# mysql> describe hashobjects;
# mysql> select * from hashobjects;
#
# Table structure for table `hashobjects`
#
CREATE TABLE hashobjects (
object_id int(11) NOT NULL,
name varchar(255) NOT NULL,
value text NOT NULL
);
insert into hashobjects values (1, 'firstname', 'Jack');
insert into hashobjects values (1, 'bgcolor', '#FFFFFF');
insert into hashobjects values (1, 'page', '<HTML><HEAD></HEAD>
<BODY BGCOLOR="<TMPL_VAR NAME="bgcolor">">
Welcome, <TMPL_VAR NAME="firstname">!<P>
</BODY></HTML>');
Re: Get a hash object from a database with DBI for use with HTML::Template
by Cine (Friar) on Aug 04, 2002 at 14:36 UTC
|
A couple of comments on your code:
- You should use placeholders for values where possible for many reasons, including security, problems with quoting etc. That is you should change
my $stmt = "SELECT * FROM $table WHERE $key_col = '$key_val'";
my $sth = $dbhandle->prepare( $stmt );
$sth->execute();
into
my $stmt = "SELECT * FROM $table WHERE $key_col = ?";
my $sth = $dbhandle->prepare( $stmt );
$sth->execute($key_val);
- You should errorcheck your database stuff. Add a || die $DBI::errstr to all your DBI->connect, $dbh->prepare and $sth->execute.
- You only need two row, so only get those. That is change:
my $ref = $sth->fetchall_arrayref();
$sth->finish();
return map { $_->[$name_col] => $_->[$val_col] } @$ref;
into
my $ref = $sth->fetchall_hashref([$name_col,$val_col]);
$sth->finish();
return map {@$_} @$ref;
- print "Content-Type: text/html\n\n"; is wrong. The HTTP standard says it should be \r\n\r\n. However most browsers dont care ;)
- Dont use a & in a sub call without needing it. It overrides any prototype you may add at a later time to make perl check the number/type of params to get_hash.
- Dont overgeneralize. If you only use get_hash for this specific purpose, then fill in the tablename/colname and select only the two rows you need instead of *.
T
I
M
T
O
W
T
D
I | [reply] [d/l] [select] |
|
my $dbh = DBI->connect(
"dbi:mysql:mydb", "user", "pass",
{RaiseError => 1},
);
"Dont use a & in a sub call without needing it. It overrides any prototype you may add ..."
Perl prototypes really are not very good, IMHO. I prefer to avoid them, opting for
good documentation instead. Check out (tye)Re: A question of style for
some good info.
jeffa
L-LL-L--L-LL-L--L-LL-L--
-R--R-RR-R--R-RR-R--R-RR
B--B--B--B--B--B--B--B--
H---H---H---H---H---H---
(the triplet paradiddle with high-hat)
| [reply] [d/l] |
Re: Get a hash object from a database with DBI for use with HTML::Template
by dug (Chaplain) on Aug 04, 2002 at 23:19 UTC
|
Warning: this reads the entire data set into memory, so be careful not to read huge amounts of data.
You can avoid this by creating other problems :)
Inside of get_hash(),
my $sth = $dbhandle->prepare( $stmt );
can become
my $sth = $dbhandle->prepare( $stmt,
{"mysql_use_result" => 1}
);
This will block other processes, and you will need to roll throught the complete dataset with fetchall_* or fetchrow_* and friends to avoid future errors on the handle unless you want to set up a brand new database for each query. It does seem to make sense some places, however.
dug | [reply] [d/l] [select] |
|
|