Shell Programming and Scripting

vgersh99,

Thanks very much for the reply. It almost worked; the nawk choked, however as it found some instances of other spurious new lines embedded in the fields. Now, the problem needs to be formulated slightly differently:

with the original set of data, find any spurious new line characters (not necessarily anchored to "^|"). The only legitimate new lines are the ones that immediately precede the primary key integer field (3110103087 and the like, having 10 digits)

Awaiting a slightly modified solution

A Shell method. Keep building the record until we have 24 pipe delimiters.
Allows for more than one extra line break in any one record.

Code:

#!/bin/ksh
saveline=""
cat filename.txt | while read line
do
        line="${saveline}${line}"
        count=`echo "${line}"|tr -cd '|'|wc -c` # Number of pipes
        if [ $count -ne 24 ]
        then
                saveline="${line}"
                continue
        else
                saveline=""
                echo "${line}"
        fi
done

Ps. This would be better in awk !

Code:

nawk -F'|' '$1~"^[0-9][0-9]*$" && length($1)==10 {printf("%c%s", (FNR==1)?"":ORS,$0);next}{printf $0}' myFile

Moderator's Comments:

Please use code tags when posting data and code samples!

gersh99,

upon further examination of the input file I have encountered spurious new lines that were not anchored to pattern "^|". In such cases, your "nawk" one-liner choked with the following message:

awk: 0602-565 There are not enough parameters in printf statement Both Eyes VA TA Dilate: 1%M Dilate: 2.5%N Contacts: previous Indication: Exam Eye [V72.0]|N||15589218|2011.11.1611:01:16:436||||0.
The input line number is 1742.
The source line number is 1.

The task at hand is to find (and replace) any spurious new lines NOT IMMEDIATELY preceding the integer 10-digit primary key

Please help.

lemele

Assuming that there are no embedded pipes and all the records have the same number of "|"s, you can fix the file much easier with awk or even sed.

binlib,

thanks for your comment; can you suggest a concrete solution to the problem at hand?

Thanks

Does my Shell version work?
If it does, this will give the "awk" programmers a program design.