Shell Programming and Scripting

Hi, Unix experts,
I have a problem about one file. my source file as following:
123, accounting, university street, 23456
124, financing, center street, b suit, 23457
125, information, college street, 23458
above file, 'center street, b suit' should be one field, my file contains more than 50 columns.
target should be
123, accounting, university street, 23456
124, financing, center street b suit, 23457
125, information, college street, 23458

how can I fix this problem? anybody can give me idea.

Thanks in advance.

This should work, but everything is hard coded:

Code:

awk -F, '{for(i=1;i<=NF;++i) {if($i==" center street") {printf "%s %s", $i, $(i+1); ++i} else {printf $i} if (i<NF) printf ", "} print ""}' infile

or simply sed:

Code:

sed 's/\(center street\),\( b suit\)/\1\2/g' infile

Hi, Kevintse,
Thanks for your replay, but my problem, the column which contains delimiter has different contents. for example.
123, accounting, university street, 23456
124, financing, center street, b suit, 23457
125, information, college street, 23458
126, computer, clars street, apt 12, 2222
target should be
123, accounting, university street, 23456
124, financing, center street b suit, 23457
125, information, college street, 23458
126, computer, clars street apt 12, 2222


Thanks

What this GNU sed line will do is capture everything EXCEPT the extra comma of the address line 2 field.

Code:

gsed 's/^\([^,]*,[^,]*,[^,]*\),\([^,]*,[^$]*$\)/\1\2/'

An awk solution (you need to adapt it to your field number):

Code:

awk -F\, 'NF==5{$3=$3""$4;$4="";sub(/,,/,",")}1' OFS="," infile

Code:

perl -ne 's/(([^,]+,\s+){2})([^,]+),/$1$3/; print ' input_file

If the total number of comma's in a line are less than or equal to 4 then try..

Code:

sed -e 's/,//3' -e s'/\([a-z]\+\) \([0-9]\+\)/\1,\2/' inputfile