Trying to extract column names from an SQL query

Question

I've got a very large and nasty set of queries from which I want to extract column names, they will all be of the format <table>.<column>

I'm new to regex and can't seem to see where my pattern match is failing, it won't return anything.

# Read file in and parse out all column names, they must be of the
# form <table>.<column> 
$data_file="C:\\VW_DEP_MTHLY_PROJTD_UNPROJTD_STK_FACT.sql";
open SQL_FILE, $data_file or die "Could not open file: $data_file";

while (<SQL_FILE>) {
    if((/ \w+\.\w+/)) {
        print $1;
    }
}

close SQL_FILE;

You can imagine part of the query like this:

    SELECT PHARMACY_FACT.REC_TYP_ID,
         PHARMACY_FACT.PACK_ID,
         ("VARCHAR"(DATE_PART('YEAR'::"VARCHAR", ASSET.MTHLY_CYC_END_DT)) || LPAD("VARCHAR"(ASSET.IMS_CYC_WK_NBR), 2, '0'::"VARCHAR")) AS IMS_CYC_WK_NBR,
         ASSET.WKLY_CYC_END_DT,
         PHARMACY_FACT.TAX_PCT_RT,
         CASE WHEN ((CASE WHEN (PHARMACY_FACT.INDIR_PUR_AMT NOTNULL) THEN PHARMACY_FACT.INDIR_PUR_AMT WHEN (0 NOTNULL) THEN '0'::"NUMERIC"
    ELSE NULL::"NUMERIC"
    END + CASE WHEN (PHARMACY_FACT.DIR_PUR_AMT NOTNULL) THEN PHARMACY_FACT.DIR_PUR_AMT WHEN (0 NOTNULL) THEN '0'::"NUMERIC"
    ELSE NULL::"NUMERIC"
    END) <> '0'::"NUMERIC") THEN (CASE WHEN (PHARMACY_FACT.INDIR_PUR_AMT NOTNULL) THEN PHARMACY_FACT.INDIR_PUR_AMT WHEN (0 NOTNULL) THEN '0'::"NUMERIC"
    ELSE NULL::"NUMERIC"
    END + CASE WHEN (PHARMACY_FACT.DIR_PUR_AMT NOTNULL) THEN PHARMACY_FACT.DIR_PUR_AMT WHEN (0 NOTNULL) THEN '0'::"NUMERIC"
    ELSE NULL::"NUMERIC"
    END)
    ELSE NULL::"NUMERIC"
    END AS UNPRJ_TOT_PUR_AMT
    ...

Answer 1

In order for $1 to be set, you need to "capture" some part of the regex by putting it in parentheses.

while (<SQL_FILE>) {
    if (/ (\w+\.\w+)/) {
        print $1;
    }
}

Looks like you had the parens outside of the match operator rather than inside it.

Answer 2

your code will find only first column name in each line. You can split each line by spaces and parse primitives

use strict;

my $data_file='/tmp/sosql';
open my $sql_file, $data_file or die "Could not open file: $data_file";

while (my $line = <$sql_file>) {
    foreach my $word (split(/\s/,$line)) {
        if ($word =~ /(\w+\.\w+)/) {
            print $1."\n";
        }
    }
}

close $sql_file;

Answer 3

You have to capture the column name in a group using parenthesis and loop on all column within eacch line:

while (<SQL_FILE>) {
    while(/(\w+\.\w+)/g) {
        print $1,"\n";
    }
}