Parse a file to remove " and replace , with ~

vineetjoshi01 · New User Joined: 06 Nov 2009 Posts: 14 Location: Singapore

Hi All,

I have a file as mentioned in below sample. I need to parse this file and remove all the quotes " and have to replace all the commas , with ~ sign.

Input: The lenght of all the records are not same.

2004-10-12,"DOUGL",1724,1,"832702","M",4960,"O","20376361",1
2004-10-12,"DOU",1723,1,"161318","M",5119,"OW","20377946",1
2004-10-12,"DOUGLPA",1724,1,"832702","M",4960,"O","20376361",1

Output: The output should be as below

2004-10-12~DOUGL~1724~1~832702~M~4960~O~20376361~1
2004-10-12~DOU~1723~1~161318~M~5119~OW~20377946~1
2004-10-12~DOUGLPA~1724~1~832702~M~4960~O~20376361~1

Regards,
Vineet

Bill Woodger · Posted: Mon Jul 23, 2012 5:49 pm

Please use the Code tags for your data.

Look at FINDREP in the manual. If I have understood correctly, it can do what you want.

vasanthz · Posted: Mon Jul 23, 2012 5:56 pm

Hello,
This may work,

Bill Woodger · Posted: Mon Jul 23, 2012 6:06 pm

Basically there, but will be simpler if all the quotes are removed first, then all the commas can be changed to tildes. C'' (a character string of no length) indicates the removal of the position occupied by that value. By default the record will be "shuffled" to the left.

vasanthz · Posted: Mon Jul 23, 2012 6:11 pm

Hello Bill,
You are right, I missed to notice that all fields are separated by , and " were expendable.

vineetjoshi01 · New User Joined: 06 Nov 2009 Posts: 14 Location: Singapore

Hi Vasanth / Bill,

Thanks for your help!!! It worked fine!!!

Regards,
Vineet

vineetjoshi01 · New User Joined: 06 Nov 2009 Posts: 14 Location: Singapore

Hi Vasanth / Bill,

Could you please help further in this case. Actually I have got comma , as data in some of the record fields. For example "DOUGL,HARON" in below record. I do not want that , to be replaced with ~. So the input will be as follows:

2004-10-12,"DOUGL,HARON",1724,1,"832702","M",4960,"O","20376361",1
2004-10-12,"DOU",1723,1,"161318","M",5119,"OW","20377946",1
2004-10-12,"DOUGLPA",1724,1,"832702","M",4960,"O","20376361",1

Output should be as follows: The , is retained which is part of data but rest of the commas are replaced with ~ sign.

2004-10-12~DOUGL,HARON~1724~1~832702~M~4960~O~20376361~1
2004-10-12~DOU~1723~1~161318~M~5119~OW~20377946~1
2004-10-12~DOUGLPA~1724~1~832702~M~4960~O~20376361~1

Thanks and regards,
Vineet

dbzTHEdinosauer · Posted: Thu Aug 02, 2012 3:09 pm

looks like you will have to go to parse using the PAIR option
and then BUILD, inserting the tilde between fields.

Bill Woodger · Posted: Thu Aug 02, 2012 4:34 pm

That'll do it. You PARSE the fields using the comma as a delimiter. Those bounded by quotes, use PAIR to preserve any commas inside. Then BUILD your output, with the tilde between each field. Your quote-protected commas will be preserved that way.

Skolusu · Posted: Thu Aug 02, 2012 10:18 pm

vineetjoshi01,

Use the following DFSORT JCL which will give you the desired results. The trick here is to use FINDREP to replace the comma to a space and then use SQZ with PAIR=QUOTE to replace the space with ~. Once that is done we use another SQZ to replace the space within the name to comma. Finally we use another FINDREP to replace the double quotes.