dbmathis has asked for the wisdom of the Perl Monks concerning the following question:

Hi,

I have been given a critical task of updating 45000 records on sensitive oracle database and although it is a simple update, I am not to familiar with writing update scripts in perl. I normally just do selects and then do data analysis

I am looking for a full script example using DBI that will update the records and keep a log of everything that gets updated.

My update statement is the following:
update
    ( 
    select t1.field1, 999 deact_status 
    from  
        schema.table1 t1
        inner join schema.table2 t2 on t1.id = t2.id
    where 
        t2.field2 = $emp_id
    ) 
set 
    field1 = deact_status

[download]
And the 45000 emp_ids exist in a flat file in the following format which is an emp_id per line.
0909099
3298729
2849872
2472972
...
...
...

[download]

I am sure its as easy as just loooping through the file one emp_id at a time and then plugging the emp_id into the update statement and running it, however I want to learn how to do this the right way at the beginning from the pros.

After all this is over, all that will really have mattered is how we treated each other.

Replies are listed 'Best First'.
Re: Oracle mass update script request.
by Gangabass (Vicar) on Jul 28, 2008 at 02:00 UTC

    You right about looping. Just use placeholders.

    So:

    #connect to DB
.......
#prepare your update statement
.......
#YOUR LOOP START
  #read line from your file (via while for example)
  #execute your statement
#YOUR LOOP END

    [download]
[reply]
[d/l]
Re: Oracle mass update script request.
by TedPride (Priest) on Jul 28, 2008 at 04:06 UTC
    Assuming your update statement is correct, given your database structure and Oracle (I'm more familiar with mySQL): 
    use DBI;
use strict;

my $user = 'user';
my $pass = 'pass';
my $db = 'database';
my $filename = 'filename.txt';

my $query = qq*
update
    ( 
    select t1.field1, 999 deact_status 
    from  
        schema.table1 t1
        inner join schema.table2 t2 on t1.id = t2.id
    where 
        t2.field2 = ?
    ) 
set 
    field1 = deact_status
*;

my ($dbh, $sth, $handle);

$dbh = DBI->connect("dbi:Oracle:$db", $user, $pass)
    ||  die "Couldn't connect to database: " . DBI->errstr;

open ($handle, $filename)
    || die "Couldn't open $filename for read.\n";

while (<$handle>) {
    chomp;
    $sth = $dbh->prepare_cached($query, {}, 1);
    $sth->execute($_);
}

$dbh->disconnect();

    [download]
    This is untested, but ought to work in theory.
[reply]
[d/l]
Re: Oracle mass update script request.
by grinder (Bishop) on Jul 28, 2008 at 10:21 UTC

    Why do you want to make life so miserable for yourself? Why not ask the database to do the update for you in one hit.

    update table1 t1
set t1.field1 = 999
where exists (
    select 1 from table2 t2 where t1.id = t2.id
)

    [download]

    Then you just have to debug your SQL. No Perl needed apart from a $db->do(''), no bugs to look for! Not to mention faster...

    • another intruder with the mooring in the heart of the Perl

[reply]
[d/l]
      grinder,

      I must be missing something. Are you saying I can update 45000 records all at once? Were does the sql get the 45000 records from when they only exist in a flat text file?

      I agree that the easiest solution is always the best solution. Can you please elaborate?

      Best Regards

      After all this is over, all that will really have mattered is how we treated each other.
[reply]
[reply]

Back to Seekers of Perl Wisdom