Question

Ok, so here’s the thing. I have note in an old sql server text format. It puts all notes for a record in one big blob of data. I need to take that blob of text and parse it out to create one row for each note entry with separate columns for timestamp, user, and note text. The only way do do this that I can think of is to use regex to locate the unix timestamp for each note and parse on that. I know that there is the split function for parsing on delimiters, but that removes the delimiter. I need to parse on \d{10} but also retain the 10 digit number. Here is some sample data.

create table test_table
(
job_number number,
notes varchar2(4000)
)

insert into test_table values
(12345, '1234567890 username notes text notes text notes text notes text 5468204562 username notes text notes text notes text notes text 1025478510 username notes text notes text notes text notes text')
(12346, '2345678901 username notes text notes text notes text notes text 1523024512 username notes text notes text notes text notes text 1578451236 username notes text notes text notes text notes text')
(12347, '2345678902 username notes text notes text notes text notes text 2365201214 username notes text notes text notes text notes text 1202154215 username notes text notes text notes text notes text')

I would like to see one record for each note to look like this.

JOB_NUMBER        DTTM    USER     NOTES_TEXT
----------    ----------  ----     ----------
12345         1234567890  USERNAME notes text notes text notes text notes text
12345         5468204562  USERNAME notes text notes text notes text notes text
12345         1025478510  USERNAME notes text notes text notes text notes text
12346         2345678901  USERNAME notes text notes text notes text notes text
12346         1523024512  USERNAME notes text notes text notes text notes text
12346         1578451236  USERNAME notes text notes text notes text notes text
12347         2345678902  USERNAME notes text notes text notes text notes text
12347         2365201214  USERNAME notes text notes text notes text notes text
12347         1202154215  USERNAME notes text notes text notes text notes text

Thank you for any help you can provide

Answer 1

Text::ParseWords can handle the quoted strings and split on comma. You can skip ahead in the input by using the flip-flop operator 1 .. /values/. This particular skip method may need to be revised.

Then it is just a matter of parsing the strings, which can be done by splitting using a lookahead assertion and then capturing the various entries in each substring. The regex in the split:

my @entries = split /(?<!^)(?=\d{10})/, $data;

has a negative lookbehind assertion to avoid matching at the start of the string ^, and a lookahead assertion to match 10 numbers. This will effectively split at the numbers and keep them.

The DATA file handle is used for demonstration, simply replace <DATA> with <> to use with argument file name.

use strict;
use warnings;
use Text::ParseWords;

my $format = "%-12s %-12s %-10s %s\n";              # format for printing
my @headers = qw(JOB_NUMBER DTTM USER NOTES_TEXT);  
printf $format, @headers;
printf $format, map "-" x length, @headers;         # print underline
while (<DATA>) {
    next while 1 .. /values/;                       # skip to data
    s/^\(|\)$//g;                                   # remove parentheses
    my ($job, $data) = quotewords('\s*,\s*',0, $_); # parse string
    my @entries = split /(?<!^)(?=\d{10})/, $data;  # split into entries
    for my $entry (@entries) {                      # parse each entry
        my ($dttm, $user, $notes) = $entry =~ /^(\d+)\s+(\S+)\s+(.*)/;
        printf $format, $job, $dttm, $user, $entry;
    }
}

__DATA__
create table test_table
(
job_number number,
notes varchar2(4000)
)

insert into test_table values
(12345, '1234567890 username notes text notes text notes text notes text 5468204562 username notes text notes text notes text notes text 1025478510 username notes text notes text notes text notes text')
(12346, '2345678901 username notes text notes text notes text notes text 1523024512 username notes text notes text notes text notes text 1578451236 username notes text notes text notes text notes text')
(12347, '2345678902 username notes text notes text notes text notes text 2365201214 username notes text notes text notes text notes text 1202154215 username notes text notes text notes text notes text')

Output:

JOB_NUMBER   DTTM         USER       NOTES_TEXT
----------   ----         ----       ----------
12345        1234567890   username   1234567890 username notes text notes text notes text notes text
12345        5468204562   username   5468204562 username notes text notes text notes text notes text
12345        1025478510   username   1025478510 username notes text notes text notes text notes text
12346        2345678901   username   2345678901 username notes text notes text notes text notes text
12346        1523024512   username   1523024512 username notes text notes text notes text notes text
12346        1578451236   username   1578451236 username notes text notes text notes text notes text
12347        2345678902   username   2345678902 username notes text notes text notes text notes text
12347        2365201214   username   2365201214 username notes text notes text notes text notes text
12347        1202154215   username   1202154215 username notes text notes text notes text notes text