Shell Programming and Scripting

Hello. I am trying to convert occurrences of 'NULL' from a datafile. The 'NULL' occurences appears at this:

|NULL| NULL|NULL| NULL|NULL| NULL|NULL| NULL|

There should be 52 fields per line.
I would like any occurrence of | NULL| or |NULL| to appear as '||'
Currently I am using this sed command to convert:

Code:

/usr/bin/sed -e '1,2d' -e '/^$/d' -e '$d' ${DATA}/in1.bcp > ${DATA}/trash.bcp
/usr/bin/sed -e 's/^.//g' -e 's/  */ /g' -e 's/|$//g' -e 's/000|/0|/g' -e 's/ |/|/g' ${DATA}/trash.bcp > ${DATA}/out1.bcp
/usr/bin/sed -e 's/|NULL|/|/g' -e 's/|*NULL|/|/g' ${DATA}/out1.bcp > ${DATA}/final.bcp

The problem is some fields come out like '| |' and the number of fields do not equal 52. They vary depending on how many occurrences where converted per line.

Thank you for your help.

Look at this from your last line:
-e 's/|NULL|/|/g' -e 's/|*NULL|/|/g'
If I understand what you want, you should be doing
-e 's/|NULL|/||/g' -e 's/|*NULL|/||/g'

Except that first command handles stuff also handled by the second command. Remember that * means zero or more.

Your code seems to be doing various things but for the problem you describe, why not just divide your sed statement into 2 commands, one that removes 'NULL's and the other that removes spaces.


sed -e 's/ //g' -e 's/NULL//g'