The problem is, even though the program is not finding the key value - it is still deleting it.
I'm trying to figure out how you are able to reach this conclusion. I had to modify the code so that I could read it (fix indentation, remove unnecessary variable declarations, and generally simplify things), but having done that, I don't see anything in the code that would be deleting hash keys unless the hash data produced at least one non-empty return from at least one database server.
I noticed that when a db query returns a row, you want to print out the row from the db together with the data from the "input_file.csv" that was used to make the query. The query is always asking for eight fields, so if there was a matching row, the returned array should have 8 elements (even if some of them are undef/null). This means that the second "if" statement in this snippet should be unnecessary:
(That's an example of how I simplified the original -- I believe this version should do the same thing as the original code, and it's a lot easier to read.) My point here is that you could remove the second "if" statement, leaving just one "if" block with 3 lines of code (a one-line "for" loop, a "print", and a "delete").if (@dbrec) { if ( @dbrec > 6 ) { $_ ||= '' for ( @dbrec ); print join( '|', @drec[0..13], @dbrec[0..7] ), "\n"; } delete $list{$key}; }
Anyway, when you say "it's not finding a match, but it's still deleting the key", is this because you are not seeing the sort of lines printed by that snippet, and you are also not seeing any lines printed from the later "for" loop?
Try stepping through the script with "perl -d" (run it under the perl debugger). Best plan would be to set a breakpoint at the line that is supposed to delete a hash key, and check the values of relevant variables when you get to that line (if you get to that line).
Adding "use Data::Dumper" to the script will also help, because that lets you run "print Dumper($hashref)" as a debugger command, so you can see what the data really looks like. I'm guessing there may be something unexpected in your "input.csv" file (or in how it actually works with the query).
One last suggestion: I think you are wasting a lot of execution time (and server resources) by preparing your query on every iteration. You would be better off with a loop structure where you prepare the query just once for each db connection, using a "?" placeholder for the variable, like this:
(update: deleted an unnecessary line of code from the second "for" loop)my $sql = <<ENDSQL; select ... from ... ... where so.value = ? ENDSQL for my $lookup ( sort keys %dbsrv ) { # ... make the connection as you do now do { warn "server $lookup: connect failed\n"; next } unless $dbh; my $sth = $dbh->prepare( $sql ); # "prepare" once do { warn "server $lookup: prepare failed\n"; next } unless $sth; for my $key (sort keys %list) { for my $rpt ( sort keys %{$list{$key}} ) { my @drec = @{$list{$key}{$rpt}}; my $phone = $drec[1]; $sth->execute( $phone ); # "execute" repeatedly my @dbrec = $sth->fetchrow_array; if ( @dbrec ) { ... } } } }
In addition to being simpler, faster and more efficient, that approach is also safer / more reliable (in case the input data that you pass to the query happens to contain an apostrophe, for instance).
update: Forgot to mention: I can understand that as you loop over different servers, you don't want the script to "die" just because a connection or a query happens to fail on one of them. But you really should add some warnings that let you know when a given server did not work as expected. I just updated the last snippet above (within a few minutes after posting it) to include such warnings: do {warn "..."; next} unless ( ... );
In reply to Re: Problem with deleting a hash key
by graff
in thread Problem with deleting a hash key
by smrobin
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |