I have an awk script that I am trying to run from the command prompt and for some reason it errors out at the gsub and sub part.

Here is what I enter on the command line

awk - f /usr/appl/riv/home/testriv/awk/riv_feed_import.awk /usr/appl/riv/home/testriv/data/eagle_ac

and the awk script part where its failing is-

{
# remove trailing spaces
sub(/[ \t]+$/, ""); # set number of placeholders for bcp
myPlaceHolders=substr(OFS_PLACE_HOLDERS,0,NUM_OUTP UT_COLS-NF-1);
# change FS to OFS
reFS = "[\\" FS "]";
gsub(reFS,OFS);
# write out the bcp formatted data feed
print FEED_CD,STAGE_TABLE,$0,myPlaceHolders > FILENAME ".bcp";
}

First of all: awk comes in several versions. The sub and gsub functions tipically belong to the enhanced '85 version. Previously I worked on an old Solaris 2.5 OS where awk invoked the original '77 version. If I had to use extended functionality I had to use nawk instead. There are more systems like that and you should check this out because your sub fubction seems perfectly right to me.

Second: To my knowledge the sub and gsub functions only operate with literal regexp's, not with parameterized ones. My way of working around this would be creating a shellscript that first writes an awk commandfile to disk (with the rexexp dynamically substituted from a variable) and then executes this file.

Third: If it's your objective here to change the fieldseparator which is performed on $0 you should do it like 'OFS = fieldseparator'. I believe your gsub tries to substitute every occurance of FS or backslash in $0 with the OFS but (g)sub only operates on strings and fieldseparators are not part of that.

Regards