View previous topic :: View next topic
|
Author |
Message |
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
Hi,
I need to find all the members of a partitioned library that have two or more words inside them, even if not in the same line.
With batch SRCHFOR this is not possible, it finds the members that have the two words in the same row.
I was wondering if there is a way with batch FILEAID(FMNMAIN?).
The alternative would be to do more extractions and make the appropriate pairings with SORT, but I have very large libraries and the execution time is too much.
Thank you |
|
Back to top |
|
|
sergeyken
Senior Member
Joined: 29 Apr 2008 Posts: 2140 Location: USA
|
|
|
|
SORT has no options to handle PDS datasets.
If your library is RECFM=FB,LRECL=80, then you can do this:
1) "unload" the library to a DSORG=PS dataset, using IEBPTPCH or other available utility
2) use SORT/ICETOOL/SYNCTOOL to process this flat temporary dataset, and to detect required original members. It is not very trivial task, but not rocket science as well.
3) SORT works extremely fast compared to other 99% of available tools.
Another option may be (with any RECFM/LRECL!):
1) use SUPERC to find the members with either of two required words, and
2) use SORT/ICETOOL/SYNCTOOL to re-process the output from SUPERC, and to detect the members with both required words.
The complexity of SORT control statements would be about the same for both options.
You can try it yourself, before asking to provide you with a ready-to-use solution. |
|
Back to top |
|
|
prino
Senior Member
Joined: 07 Feb 2009 Posts: 1314 Location: Vilnius, Lithuania
|
|
|
|
RedWolf wrote: |
Hi,
With batch SRCHFOR this is not possible, it finds the members that have the two words in the same row. |
That's bullsh*t. You can specify multiple search terms and SRCHFOR finds all members, no matter how many words are on separate lines! |
|
Back to top |
|
|
sergeyken
Senior Member
Joined: 29 Apr 2008 Posts: 2140 Location: USA
|
|
|
|
RedWolf wrote: |
Hi,
I need to find all the members of a partitioned library that have two or more words inside them, even if not in the same line.
With batch SRCHFOR this is not possible, it finds the members that have the two words in the same row.
I was wondering if there is a way with batch FILEAID(FMNMAIN?).
The alternative would be to do more extractions and make the appropriate pairings with SORT, but I have very large libraries and the execution time is too much.
Thank you |
ENOUGH THEORETICAL BLAH-BLAH-BLAH.
Bring here your examples: where your "With batch SRCHFOR this is not possible, it finds the members that have the two words in the same row"? |
|
Back to top |
|
|
Joerg.Findeisen
Senior Member
Joined: 15 Aug 2015 Posts: 1334 Location: Bamberg, Germany
|
|
|
|
It seems that FINDALL option that @prino has secretly referred to, only works for PS datasets. Maybe I am missing another keyword so far which enables this for PO as well. |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
prino wrote: |
RedWolf wrote: |
Hi,
With batch SRCHFOR this is not possible, it finds the members that have the two words in the same row. |
That's bullsh*t. You can specify multiple search terms and SRCHFOR finds all members, no matter how many words are on separate lines! |
--
ok, maybe I got some parameters wrong. I'll show you what I do and what I get as output
the contents of the members are:
MEMBER A
MEMBRO A 1 PIPPO
MEMBER B
MEMBRO B 1 PLUTO
MEMBRO C
MEMBRO C 1
MEMBRO D
MEMBRO D 1 PIPPO
MEMBRO D 1 PLUTO
I want only MEMBER D (contains the two words, in different lines),
with this search:
//SEARCH EXEC PGM=ISRSUPC,
// PARM=(SRCHCMP)
//* 'FMSTOP')
//NEWDD DD DISP=SHR,DSN=XXXXXXX.XXX.XXX
//OUTDD DD SYSOUT=*
//SYSIN DD *
SRCHFOR 'PIPPO' AND 'PLUTO'
but I get this:
A --------- STRING(S) FOUND -------------------
1 MEMBRO A 1 PIPPO
D --------- STRING(S) FOUND -------------------
1 MEMBRO D 1 PIPPO
ISRSUPC - MVS/PDF FILE/LINE/WORD/BYTE/SFOR COMPARE UTILITY- ISPF FOR z/OS
SEARCH-FOR SUMMARY SECTION SRCH DSN: U375469.LIB.INP
LINES-FOUND LINES-PROC MEMBERS-W/LNS MEMBERS-WO/LNS COMPARE-COLS LONGEST-LI
2 11 2 3 1:80 80
THE FOLLOWING PROCESS STATEMENTS (USING COLUMNS 1:72) WERE PROCESSED:
SRCHFOR 'PIPPO' 'PLUTO'
ISRS024W EXTRA DATA DETECTED AFTER NORMAL STATEMENT END. STATEMENT ACCEPTED WITH WARNING NOTIFICATION.
i tried also
SRCHFOR 'PIPPO' 'PLUTO' |
|
Back to top |
|
|
Joerg.Findeisen
Senior Member
Joined: 15 Aug 2015 Posts: 1334 Location: Bamberg, Germany
|
|
|
|
See my comment above yours, and also have a look at the allowed SRCHFOR term(s).
What RECFM;LRECL does your DSN have? |
|
Back to top |
|
|
sergeyken
Senior Member
Joined: 29 Apr 2008 Posts: 2140 Location: USA
|
|
|
|
1. Learn how to use the Code button when presenting your data and code to others.
RedWolf wrote: |
the contents of the members are:
Code: |
MEMBER A
MEMBRO A 1 PIPPO
MEMBER B
MEMBRO B 1 PLUTO
MEMBRO C
MEMBRO C 1
MEMBRO D
MEMBRO D 1 PIPPO
MEMBRO D 1 PLUTO |
I want only MEMBER D (contains the two words, in different lines),
with this search:
Code: |
//SEARCH EXEC PGM=ISRSUPC,
// PARM=(SRCHCMP)
//* 'FMSTOP')
//NEWDD DD DISP=SHR,DSN=XXXXXXX.XXX.XXX
//OUTDD DD SYSOUT=*
//SYSIN DD *
SRCHFOR 'PIPPO' AND 'PLUTO'
|
but I get this:
Code: |
A --------- STRING(S) FOUND -------------------
1 MEMBRO A 1 PIPPO
D --------- STRING(S) FOUND -------------------
1 MEMBRO D 1 PIPPO
ISRSUPC - MVS/PDF FILE/LINE/WORD/BYTE/SFOR COMPARE UTILITY- ISPF FOR z/OS
SEARCH-FOR SUMMARY SECTION SRCH DSN: U375469.LIB.INP
LINES-FOUND LINES-PROC MEMBERS-W/LNS MEMBERS-WO/LNS COMPARE-COLS LONGEST-LI
2 11 2 3 1:80 80
THE FOLLOWING PROCESS STATEMENTS (USING COLUMNS 1:72) WERE PROCESSED:
SRCHFOR 'PIPPO' 'PLUTO'
ISRS024W EXTRA DATA DETECTED AFTER NORMAL STATEMENT END. STATEMENT ACCEPTED WITH WARNING NOTIFICATION.
|
i tried also
Code: |
SRCHFOR 'PIPPO' 'PLUTO' |
|
2. Use SORT/ICETOOL/SYNCTOOL to re-process the output from SUPERC, and to detect the members with both required words. |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
sergeyken wrote: |
SORT has no options to handle PDS datasets.
If your library is RECFM=FB,LRECL=80, then you can do this:
1) "unload" the library to a DSORG=PS dataset, using IEBPTPCH or other available utility
2) use SORT/ICETOOL/SYNCTOOL to process this flat temporary dataset, and to detect required original members. It is not very trivial task, but not rocket science as well.
3) SORT works extremely fast compared to other 99% of available tools.
Another option may be (with any RECFM/LRECL!):
1) use SUPERC to find the members with either of two required words, and
2) use SORT/ICETOOL/SYNCTOOL to re-process the output from SUPERC, and to detect the members with both required words.
The complexity of SORT control statements would be about the same for both options.
You can try it yourself, before asking to provide you with a ready-to-use solution. |
--
I've already done what you say, but I have a library with hundreds of thousands of members... the processing time is long and the sequential file I get is very large. I was looking for a faster solution.
Thanks |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
Joerg.Findeisen wrote: |
See my comment above yours, and also have a look at the allowed SRCHFOR term(s).
What RECFM;LRECL does your DSN have? |
My DSN is a Partitioned Dataset, with FB 80... a classical partitioned with jcl inside. A lot of JCL! |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
sergeyken wrote: |
1. Learn how to use the Code button when presenting your data and code to others.
RedWolf wrote: |
the contents of the members are:
Code: |
MEMBER A
MEMBRO A 1 PIPPO
MEMBER B
MEMBRO B 1 PLUTO
MEMBRO C
MEMBRO C 1
MEMBRO D
MEMBRO D 1 PIPPO
MEMBRO D 1 PLUTO |
I want only MEMBER D (contains the two words, in different lines),
with this search:
Code: |
//SEARCH EXEC PGM=ISRSUPC,
// PARM=(SRCHCMP)
//* 'FMSTOP')
//NEWDD DD DISP=SHR,DSN=XXXXXXX.XXX.XXX
//OUTDD DD SYSOUT=*
//SYSIN DD *
SRCHFOR 'PIPPO' AND 'PLUTO'
|
but I get this:
Code: |
A --------- STRING(S) FOUND -------------------
1 MEMBRO A 1 PIPPO
D --------- STRING(S) FOUND -------------------
1 MEMBRO D 1 PIPPO
ISRSUPC - MVS/PDF FILE/LINE/WORD/BYTE/SFOR COMPARE UTILITY- ISPF FOR z/OS
SEARCH-FOR SUMMARY SECTION SRCH DSN: U375469.LIB.INP
LINES-FOUND LINES-PROC MEMBERS-W/LNS MEMBERS-WO/LNS COMPARE-COLS LONGEST-LI
2 11 2 3 1:80 80
THE FOLLOWING PROCESS STATEMENTS (USING COLUMNS 1:72) WERE PROCESSED:
SRCHFOR 'PIPPO' 'PLUTO'
ISRS024W EXTRA DATA DETECTED AFTER NORMAL STATEMENT END. STATEMENT ACCEPTED WITH WARNING NOTIFICATION.
|
i tried also
Code: |
SRCHFOR 'PIPPO' 'PLUTO' |
|
2. Use SORT/ICETOOL/SYNCTOOL to re-process the output from SUPERC, and to detect the members with both required words. |
Thanks for your help in providing the correct answer. I have already applied the solution you proposed, but I was looking for something faster in processing |
|
Back to top |
|
|
sergeyken
Senior Member
Joined: 29 Apr 2008 Posts: 2140 Location: USA
|
|
|
|
RedWolf wrote: |
sergeyken wrote: |
SORT has no options to handle PDS datasets.
If your library is RECFM=FB,LRECL=80, then you can do this:
1) "unload" the library to a DSORG=PS dataset, using IEBPTPCH or other available utility
2) use SORT/ICETOOL/SYNCTOOL to process this flat temporary dataset, and to detect required original members. It is not very trivial task, but not rocket science as well.
3) SORT works extremely fast compared to other 99% of available tools.
Another option may be (with any RECFM/LRECL!):
1) use SUPERC to find the members with either of two required words, and
2) use SORT/ICETOOL/SYNCTOOL to re-process the output from SUPERC, and to detect the members with both required words.
The complexity of SORT control statements would be about the same for both options.
You can try it yourself, before asking to provide you with a ready-to-use solution. |
--
I've already done what you say, but I have a library with hundreds of thousands of members... the processing time is long and the sequential file I get is very large. I was looking for a faster solution.
Thanks |
I doubt you have ever tried to process SUPERC output with SORT utility. I did it many times, and it works extremely fast (compared to any other approach).
Please, present here the code you have tried. Otherwise I have to consider you as a liar. |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
prino wrote: |
RedWolf wrote: |
Hi,
With batch SRCHFOR this is not possible, it finds the members that have the two words in the same row. |
That's bullsh*t. You can specify multiple search terms and SRCHFOR finds all members, no matter how many words are on separate lines! |
SRCHFOR finds all members that contain at least one word, I look for members that contain ALL the words of the search, even on different lines within the member. I have a very large patitioned as input. Do you know any parameters that allow you to do what I would like to do?
Grazie |
|
Back to top |
|
|
Joerg.Findeisen
Senior Member
Joined: 15 Aug 2015 Posts: 1334 Location: Bamberg, Germany
|
|
|
|
When you have more than 32k members in that DSN, I would opt for @sergeyken's approach using SUPERC and Sort. If it's less, use the IEBPTPCH method, followed by a Sort. When you STFW you will find some samples how to do. |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
sergeyken wrote: |
RedWolf wrote: |
sergeyken wrote: |
SORT has no options to handle PDS datasets.
If your library is RECFM=FB,LRECL=80, then you can do this:
1) "unload" the library to a DSORG=PS dataset, using IEBPTPCH or other available utility
2) use SORT/ICETOOL/SYNCTOOL to process this flat temporary dataset, and to detect required original members. It is not very trivial task, but not rocket science as well.
3) SORT works extremely fast compared to other 99% of available tools.
Another option may be (with any RECFM/LRECL!):
1) use SUPERC to find the members with either of two required words, and
2) use SORT/ICETOOL/SYNCTOOL to re-process the output from SUPERC, and to detect the members with both required words.
The complexity of SORT control statements would be about the same for both options.
You can try it yourself, before asking to provide you with a ready-to-use solution. |
--
I've already done what you say, but I have a library with hundreds of thousands of members... the processing time is long and the sequential file I get is very large. I was looking for a faster solution.
Thanks |
I doubt you have ever tried to process SUPERC output with SORT utility. I did it many times, and it works extremely fast (compared to any other approach).
Please, present here the code you have tried. Otherwise I have to consider you as a liar. |
Scusami se ti scrivo in italiano, ma voglio essere chiaro. Ho 61 anni, lavoro nel mondo Host IBM da 40. Ho fatto una domanda perché non trovo risposta nemmeno su chatgpt. Non devo e non voglio dimostrare a nessuno di essere in grado di saper usare IEBPTPCH ed un paio di sort. Se si ha una risposta alla mia domanda, posta con gentilezza, la si da... altrimenti non si mette in dubbio la mia professionalità. Grazie |
|
Back to top |
|
|
RedWolf
New User
Joined: 07 Feb 2024 Posts: 8 Location: Italy
|
|
|
|
Joerg.Findeisen wrote: |
When you have more than 32k members in that DSN, I would opt for @sergeyken's approach using SUPERC and Sort. If it's less, use the IEBPTPCH method, followed by a Sort. When you STFW you will find some samples how to do. |
I will continue to use IEBPTPCH and SORT. Thanks a lot for the answer. |
|
Back to top |
|
|
Jim Evans
New User
Joined: 19 Sep 2024 Posts: 2 Location: UK
|
|
|
|
If you're licensed for SELCOPY then here's a decent solution.
SELCOPY is able to directly process the whole library using BPAM, so it's very fast ...
Code: |
//SC EXEC PGM=SELCOPY
//MYLIB DD DISP=SHR,DSN=MY.LIBRARY.WITH.MANY.MEMBERS
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
option worklen = 33333 * Allows records up to LRECL=32K.
read MYLIB dirdata * DirData means use BPAM access.
if dir * Is it a directory rec?
then pos 33001 = 8 at 1 * Save new member name.
then @w1=1 * Reset switch.
then @w2=1 * Reset switch.
then goto get * Bypass.
if pos any = 'PIPPO' !then @w1=2 * Indicate found.
if pos any = 'PLUTO' !then @w2=2 * Indicate found.
if @w1 = 2 !and @w2 = 2 * Both strings found?
then print from 33001 len=8 * Print the member name.
then flag eomemb * Bypass rest of this member.
* SELCOPY automatically loops back to the 1st statement.
/*
|
[/code] |
|
Back to top |
|
|
Jim Evans
New User
Joined: 19 Sep 2024 Posts: 2 Location: UK
|
|
|
|
Another BPAM solution using FileKit (formerly SELCOPYi - also from Compute (Bridgend) Ltd) ...
Code: |
//FK EXEC PGM=FILEKITB
//MYLIB DD DISP=SHR,DSN=MY.LIBRARY.WITH.MANY.MEMBERS
//SDEPRINT DD SYSOUT=*
//SYSPRINT DD SYSOUT=*
//SDEIN DD *
macro LIBSCAN
/*
//LIBSCAN DD *
/* REXX */
'FileIO MYLIB openread library' /* LIBRARY means use BPAM */
'FileIO MYLIB option MEMBER' /* Sets MYLIB.1.MEMBER */
/* Select only records containing EITHER string */
'FileIO MYLIB where (record << "PIPPO" or record << "PLUTO")'
do forever
W1=0; W2=0
'FileIO MYLIB sread MyRec'
if rc >= 4 then leave /* End of Library? */
do i = 1 to MyRec.0
if W1 = 0 & pos('PIPPO',MyRec.i) > 0 then W1 = 1
if W2 = 0 & pos('PLUTO',MyRec.i) > 0 then W2 = 1
if W1 = 1 & W2 = 1 then /* Got both words? */
do; say MYLIB.1.Member /* Print the member name */
leave
end
end
end
/*
|
|
|
Back to top |
|
|
|