Shell Programming and Scripting

Hi all,
i need to write a script utilizing either awk or sed to allow me to read in two values off a list file (i will probably need a delimiter value there as well).
i.e. table_name, nDays. Any ideas as to how i should start?

Thanks in Advance,
Chris

P.S. Please forgive me if none of this makes sense, this is my first actual piece of code writing in unix

The command line version of the awk tool looks something like this:

awk -F, '<pattern_search> { <action_if_search_found> } END { <final_action_after_awk_is_finished> }' <your_list_file_here>

The -F, is your delimiter, a comma. Please don't put any spaces between the F and the comma.

Since awk runs the pattern on each "record" The pattern search looks something like this..

$1 ~ /word/

Meaning, if the first field contains the string "word". All regexes must be contained within the two forward slashes. $2 and $3 are fields two and three. And then finally, in the END brackets, you can place a print command, or you can leave the END out all together if you don't need it.

If you're trying to display each record line by line (for all that matches the search pattern). Then after the pattern search, within the brackets, you need to put something like { print $1 }..

Meaning, if the pattern is found in the first field... print that first field.
If the pattern does not match, then the action will be ignored.

Also, try typing "man awk" for the unix manual.

Can u plz paste a part of the file and what output do u want ??

the list file itself just contains table names. The goal is to have it include a table name and a number that will be passed through a variable and parsed through a query.

Sample: table_name1 30
table_name2 40
and etc

the number represents the number of days from today's date that will be exported and archived from said table name.

i tested this so far for the awk: awk 'BEGIN { RS = " " } { print $1 }' archiveHistTbl.list

Original Script:
TBL_HIST_LIST=$JOB_HOME/archiveHistTbl.list
TBLLIST=$JOB_HOME/archiveTbl.list
DATADIR=/tmp
TODAY=`date +%Y%m%d`

# Set path
PATH="${JOB_HOME}:$HOME/Dear/jobs:${PATH}"

# Set date to delete from
DELDATE=`shift_date.ksh $TODAY -$1`

# For each table, delete based on dt_last_upd_hist
for tbl in `cat $TBL_HIST_LIST`
do
$ORACLE_HOME/bin/exp $DBO_USER/$DBO_PASS FILE=$DATADIR/$tbl-${DELDATE}.exp TABLE
S=$tbl QUERY=\"where dt_last_upd_hist \< to_date\(\'$DELDATE\', \'DD-MON-YYYY\'\
)\"
$ORACLE_HOME/bin/sqlplus $DBO_USER/$DBO_PASS <<!
delete from $tbl where dt_last_upd_hist < to_date('$DELDATE', 'DD-MON-YYYY');
/
!
end
done

You will need to change the "for" loop into a "while read" loop, e.g...

Code:

while read tbl days
do
  DELDATE=`shift_date.ksh $TODAY -$days`
  :
  : exp then sqlplus
  :
done < cat $TBL_HIST_LIST

Incidentally, the delete where clause can be simplified to:

Code:

where dt_last_upd_hist < sysdate - $days;

Ygor,

would i be able to use the awk script to read the 2nd column of data from the list in the code you wrote?

oh also, how would i make it so that if the 2nd column had an empty value i could set a default value of 30 to be inputted?