to read particular files where labels have been set

dilipan · New User Joined: 09 Mar 2011 Posts: 5 Location: india

i have thousands of record in a ps.i want to read only the particular records where labels have been set.

prino · Posted: Mon Mar 14, 2011 2:47 pm

Records in a PS do not have labels, problem solved.

dilipan · New User Joined: 09 Mar 2011 Posts: 5 Location: india

thank you...

dick scherrer · Posted: Mon Mar 14, 2011 7:50 pm

Hello,

Now, if you post a bit of sample input data and explain the output you want from this sample data, someone may have a suggestion. . .

dilipan · New User Joined: 09 Mar 2011 Posts: 5 Location: india

.A 001
000002 002
.B 003
000004 004
000005 005
.D 006
000007 007
000008 008
.C 009
000010 010

i want to retreive only the record which i have set label like .a,.b,.c,.d

dick scherrer · Posted: Tue Mar 15, 2011 9:50 am

Hello,

Read the file.

If the first position of the record is a period (.), do whatever otherwise read the next record.

If this does not answer what you want to do, show the sample output from this sample input.

Also, is there some reason rexx has been chosen?

Pedro · Posted: Tue Mar 15, 2011 9:57 am

> records where labels have been set.

It looks very much like the labels are only in the editor session... please clarify if that is the environment.

note: the labels do not actually exist in the dataset. That is, you cannot save the dataset with the labels.

To find editor labels, use an editor macro with:

dilipan · New User Joined: 09 Mar 2011 Posts: 5 Location: india

hi
i have attached the input what i am given and also attached output what i need.

dick scherrer · Posted: Tue Mar 15, 2011 8:01 pm

Hello,

You need to re-post your data using copy/paste and the "Code" tag (as Pedro has done). There is no reason to post "screenshots" to show 3270-screen data. . .

Then explain how the "output" was determined. It is surely not obvious line 000002 was chosen/generated/whatever. . .

Kjeld · Posted: Tue Mar 15, 2011 11:06 pm

The attached screenshots are in reverse order. Top one is the output, bottom one the input.

This presumes that you want all lines in your edit session which have no prefix labels deleted, isn't that so?

Pedro has already posted the logic to test if a label is set or not...

So the process is: Open the dataset for edit, mark the lines to be retained with labels (You might even define block labels). Run the macro. Save the dataset. You can repeat the labeling and macro execution until satisfied.
If you accidently delete records, you have to use the UNDO command, or possibly cancel edit and start over.

Pedro · Posted: Tue Mar 15, 2011 11:15 pm

I suggest the following:

1. add your labels
2. EXCLUDE ALL
3. run macro that locates all labels
4. DELETE ALL X

You can add 2 and 4 to the sample macro.

NOTE: the initial post only mentioned reading particular lines, but the sample output shows the deletion of many records. You need to re-evaluate whether you really want to delete records.

dick scherrer · Posted: Wed Mar 16, 2011 12:46 am

Thanks Kjeld,

daveporcelan · Posted: Wed Mar 16, 2011 1:18 am

This is the typical thread lately.

Topic Starter gives a vauge and incomplete requirement.

The experts swoop in with thier crystal balls and take guesses as to the real requirement. They ask for more information in a effort to get a better understanding of the requirement.

Topic starter gives little tidbits of semi-useful information, usually ignoring the questions for useful information from the experts.

Topic starter looses interest while experts continue with questions and solutions.

Now to this case:

Why are you using labels to identify the lines you want to save?

It is my understanding that labels only stay for the current edit session.
With this being the case, you have one shot to identify your data out of thousands of lines.

It may work better to overlay the zero in position 1 with a K to mark the rows you want to keep.

You can then use commands like Pedro mentioned above.

You can then change all the Ks back to zero afterwards.

I think this is yet another case of the Topic starter, coming up with a ill fated solution, and trying to find a way to make it work.

The forum is much better when the real and complete requirement is clearly stated, so valuable suggestions can be given.

Pedro · Posted: Wed Mar 16, 2011 7:38 pm

Nic Clouston · Posted: Thu Mar 17, 2011 12:21 am

No - move it but pm the OP to say it has been moved and why. This could help them to post correctly next time they haven't googled, read the manuals, read their output messages etc.