Shell Programming and Scripting

hello there.
I have a script that loads data files into ORACLE tables.
I had to then add an awk line to take any zeros out of any fields prior to load which works perfectly well on its own but I need to know how to corrrctly put it into the script. eg.

#echo " "
#echo "Checking $datafile for zeros and null them where found.."
#echo " "
# if test -r $datafile
test -r $datafile
#then
newdatafile=`nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}}1' $datafile`
chmod 777 $newdatafile;
echo " "
echo " Modified Datafile is now:= $newdatafile"
# else
# echo " "
# echo "**** `..there was an error modifying the $datafile2 datafile..."
# echo " "
echo "**** datafile = " $newdatafile
#cat " "
#exit 1

..as you can see I have the datafile [which works fine when there are no zeros in the file being loaded], I pass it through the awk script but this is where I need help. I need to know how I rename the datafile to a new name/variable name which can then be used to load into ORACLE further down the script.

thanks for looking..
Steven

If I understand the question you need to redirect the output of the awk program to another file:

Code:

 nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}}1' $datafile > new_file

Regards

yes I think you understand right its just I'm not using it correctly. When I do that, then try and call the 'newfile' later down the script it doesnt recognise it as the midified version of the original file and therefore does not load it in.

Im doing:

nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}}1' $datafile > newdatafile;
echo " Modified Datafile is now:= $newdatafile"

(the line above is blank in the log file)
...then, when trying to loadinto ORACLE..

echo "Running SQL Loader now..."
echo "**** `date +%H:%M:%S` - Loading file: $newdatafile into <table name> "

sqlldr userid=$user_id \
control=$SU_TOP/ctl/CARPARK.ctl \
log=$APPLCSF/log/CARPARK.log \
bad=$APPLCSF/log/CARPARK.bad \
data=$newdatafile \
errors=1000

LRM-00116: syntax error at 'errors' following '=' .. is shown as there's no data file..

much appreciate all comments..
Steven

Is the format of the newdatafile the same as the original file?
Do you have fields in your table that don't accept empty values?

I hope someone who's familiar with this stuff can give you the right solution.

Regards

Yes, the format of the newdatafile IS the same as the original - all Ive done is nulled a few comma-separated fields in the file.
Also, as I created the table where the file is being loaded into I made sure that EVERY field accepts null values.

Thanks for the comments - have I got this in the right forum do you think?

Steven

well, inadvertantly unix.com forum still found the answer!

Whilst reading others' posts I found when using my nawk command instead of the '1' at the end of the find and replace line, Ive used 'print' which works great!

eg.

was:- nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}}1' SLsTester_DataFile.csv > SLsOUtputFile.csv;

and now this works:

nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}} {print}' $datafile > newoutputfile.csv;
newdatafile=newoutputfile.csv;
# echo " "
echo " Modified Datafile is now:= $newdatafile"....

thanks to Franklin52 for suggestions and all for looking...

This 2 commands should give the same output:

Code:

nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}}1'

nawk 'BEGIN{FS=OFS=","}{for(i=1;i<=NF;i++){if($i=="0"){$i=""}}}{print}'

Regards