Possible to skip records until a char string found?

Div Grad · New User Joined: 08 Apr 2005 Posts: 45

I have a file that breaks into two parts, I want to drop the first part and keep only the second. I can recognize where the division in the file is by a unique character string showing up in a certain column. When I see that string, for the first time, in the specified column in the file I want to keep that line and all following lines. I will want to drop all the prior lines.

Say I had an 80 byte file where I wanted to keep all records once I found the string 'keep now' in column 1. I would want this

Skolusu · Posted: Fri Mar 14, 2008 1:14 am

Div gad,

Try this DFSORT job

gcicchet · Posted: Fri Mar 14, 2008 4:07 am

Hi,
here is a FILEAID that will do the trick as well

Div Grad · New User Joined: 08 Apr 2005 Posts: 45

Skolusu - Thanks your solution worked perfectly. Now I'm greedy and want more. I see how the numbering is being put into columns 81 through 89 to count occurences of the keyword and then the INCLUDE portion of the statement then only keeps records where the count is 1 or more.

The file I am trying to break apart has say five sections. I know where a section starts by the presence of a keyword in column 1. This section ends and the next section starts when the new keyword is encountered. I tried expanding your example to add more columns of count data so that I could then break apart the file by looking at the count columns. I tried for an hour or so to code the OVERLAY statements but couldn't do it.

I want a file that looks like this

Skolusu · Posted: Fri Mar 14, 2008 10:35 pm

Div Grad,

Do you need counts just so that want to break the contents ? you dont have to add a count for each key, you can do it all with 1 count itself. Let us know what exactly you want to do. Show us examples of sample input and output datasets along with DCb properties.

Div Grad · New User Joined: 08 Apr 2005 Posts: 45

Skolusu - Yes I just want to break the file into parts, the counts themselves aren't important other than as an aid in breaking up the file into bits.

What I'm trying to break apart is a print formatted listing of a program written in IDEAL. The listing comes out with the sections always in the following order:

IDENTIFICATION
RESOURCES
ENVIRONMENT
PARAMETER
WORKING DATA
PROCEDURE

I know when I've hit the start of a section when I see the keyword in column one. I know I've hit the end of the section, and the start of the next, when I see the next keyword show up in column 1.

I'm working with a file formatted for printing, with page headers at the top of each page showing the keyword (section name) so I can encounter one individual keyword many times in a row before hitting the next new keyword. For example if the parameter section was six pages long then I'd know where the parameter section started by the first occurance of the word 'parameter' but then I would see five more occurances of this word before I finally hit the words 'working data' and knew the next section started.

So I would want this one file:

Skolusu · Posted: Sat Mar 15, 2008 1:06 am

Div Grad,

The following DFSORT/ICETOOL jcl will give you the desired results

Div Grad · New User Joined: 08 Apr 2005 Posts: 45

For some reason this is not working for me. I finally resorted to cut/pasting in the example given and running that unaltered but I still get unexpected results. I get the following outputs:

TOOLMSG

Skolusu · Posted: Tue Mar 18, 2008 1:23 am

Div Grad,

My Job works fine with DFSORT. You are using Syncsort. DFSORT and Syncsort are competitive products. I'm a DFSORT developer. I'm happy to answer questions on DFSORT and DFSORT's ICETOOL, but I don't answer questions on Syncsort.

superk · Posted: Tue Mar 18, 2008 1:38 am

Div Grad, what changed? Your other posts clearly showed that you have or had DFSORT.

Div Grad · New User Joined: 08 Apr 2005 Posts: 45

Oops, I recently changed jobs and have moved from a dfsort shop to a syncsort shop without realizing it. I'd specify icetool in the jcl but synctool was being swapped in without my realizing it. I generally just look at the output file and not the other files generated by the run.

Sorry Everyone!

dick scherrer · Posted: Tue Mar 18, 2008 3:20 am

Hello,

As the programs are written in IDEAL, you could use the IDEAL Source Transport utility and create a sequential file of the program.

You would then not need to be concerned with headers or other formatting considerations.

superk · Posted: Tue Mar 18, 2008 8:54 pm