topic Re: Info needed from input dataset wraps onto multiple lines. in SAS Programming

Info needed from input dataset wraps onto multiple lines.

a079011 — Mon, 12 Sep 2016 18:22:02 GMT

I am reading in a dataset where the info I need to parse and work with wraps over to multiple lines. This does not happen on every line and may extend to 2 or more lines. Here is what I am working with. (I have truncated my data to make it easier to read)

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN

Status(Lost control)

User(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN

MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN

MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/MAIN State

Jobno(43228) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN STATE FAILED

User(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5

Jobno(59568) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/MAIN

Jobno(7800) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCMSIS47 KPPSCDPP.QA009/DPPSC5@.1308642

Jobno(56156) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCMSIS47 KPPSCDPP.QA/DPPSC5@.1308642/MAIN State

Jobno(9672) User(SCMSIS47) Host(SCMSIS47)

Every time “MgrMsg:” appears in cols 2 – 8 and on the next line cols 2-8 are blank. I need that line, starting at column 10 to append to the end of the line that has the MgrMsg. Desired out put would look like this

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN Status(Lost control) User(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN

MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN

MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/MAIN State Jobno(43228) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN STATE FAILED User(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5 Jobno(59568) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/MAIN Jobno(7800) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCMSIS47 KPPSCDPP.QA009/DPPSC5@.1308642 Jobno(56156) User(SCMSIS47) Host(SCMSIS47)

MgrMsg: SCMSIS47 KPPSCDPP.QA/DPPSC5@.1308642/MAIN State Jobno(9672) User(SCMSIS47) Host(SCMSIS47) Jobno(67160)

OPERCMD: BY OPSMCB01 SPINLOG

Re: Info needed from input dataset wraps onto multiple lines.

Shmuel — Mon, 12 Sep 2016 19:23:10 GMT

data want (keep=out_line);
length out_line $200; /* max length expected for a line on output */
infile cards truncover;
input @2 a_line $100. ;
test = substr(a_line,1,7);
if test = "MgrMsg:" then do;
out_line = a_line;
input @2 a_line $100. ;
test = substr(a_line,1,7);
do while (test = ' ');
out_line = catx(' ',trim(out_line), trim(a_line2);
input @2 a_line $100. ;
test = substr(a_line,1,7);
end;
output;
out_line = a_line;
end;
else do;
out_line = a_line;
output;
end;

return;
/*----------*/
cards
... enter here your input data ...
run;

Re: Info needed from input dataset wraps onto multiple lines.

Shmuel — Mon, 12 Sep 2016 19:25:26 GMT

on catx function line it should be written a_line instead a_line2

Re: Info needed from input dataset wraps onto multiple lines.

Astounding — Mon, 12 Sep 2016 20:50:44 GMT

This is untested, but should be OK:

data want;

infile rawdata end=nomore;

length message $ 1000;

retain message;

input;

if _n_=1 then message = _infile_;

else do;

if left(_infile_) in : ("MgrMsg:", "APPLMGR:") then do;

output;

message = _infile_;

end;

else message = catx(message, _infile_);

end;

if nomore then output;

run;

The END= option on the INFILE statement probably won't work with a CARDS statement ... INFILE would be more appropriate here.

Re: Info needed from input dataset wraps onto multiple lines.

a079011 — Tue, 13 Sep 2016 14:29:52 GMT

Thank you for the reply and the time you spent on this. It is not producing the output I hoped. First run only produced the "out_line" var and this var does not include all the data.

PROC PRINT

The SAS System

Obs out_line

1 MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST

2 PPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

3 MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO

4 MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/

5 MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST

6 PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

7 MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/

8 MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/

So I edited "DATA want(keep=out_line);" to just "DATA want;" This out contained the data I need so I thought I would just able to join out_line and a_line.

The SAS System 10:23 Tuesday

Obs out_line a_line

1 MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST User(SCPPSB38) Host(scppsb38)

2 PPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed PPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

3 MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO

4 MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/ Jobno(43228) User(SCMSIS47) Host(SCMSIS

5 MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST User(SCPPSB38) Host(scppsb38)

6 PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

7 MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/ Jobno(59568) User(SCMSIS47) Host(SCMSIS

8 MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/ Jobno(7800) User(SCMSIS47) Host(SCMSIS4

But when I inject more than one line below the MgrMsg that I also need to caput it is not grabbing that line.

The SAS System 10:27 Tuesda

Obs out_line a_line

1 MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST *** FIRST LINE NEEDED *****

2 *** MAY OR MAY NOT BE ANOTHER ***** *** MAY OR MAY NOT BE ANOTHER *****

3 PPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed PPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

4 MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO

5 MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/ Jobno(43228) User(SCMSIS47) Host(SCMSIS

6 MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST User(SCPPSB38) Host(scppsb38)

7 PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

8 MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/ Jobno(59568) User(SCMSIS47) Host(SCMSIS

9 MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/ Jobno(7800) User(SCMSIS47) Host(SCMSIS4

I think this code will give me a good start, just needs some tweaking. Thanks again !!!

Re: Info needed from input dataset wraps onto multiple lines.

Shmuel — Tue, 13 Sep 2016 14:47:29 GMT

As we need concatenate input lines to create the out line - I missed a retain statement.

Here is the code that might work: (NOTE - the last line(s) may be truncated as infile cards doesn't turn end of input flag)

data want (keep=out_line);
length out_line $200; /* max length expected for a line on output */

retain out_line;
infile cards truncover;
input @2 a_line $100. ;
test = substr(a_line,1,7);
if test = "MgrMsg:" then do;
out_line = a_line;
input @2 a_line $100. ;
test = substr(a_line,1,7);
do while (test = ' ');
out_line = catx(' ',trim(out_line), trim(a_line2);
input @2 a_line $100. ;
test = substr(a_line,1,7);
end;
output;
out_line = a_line;
end;
else do;
out_line = a_line;
output;
end;

return;
/*----------*/
cards
... enter here your input data ...

... add at the end a line with End Of Input indication ...
run;

Re: Info needed from input dataset wraps onto multiple lines.

a079011 — Mon, 19 Sep 2016 14:16:58 GMT

Thank you again, it is still not quite what I need. It is soooo close but it is dropping a line. The problem with the input data is it there may be zero, one or tw lines that follow that need to be included. Here is the code;

OPTION NOCENTER;
DATA want(keep=out_line);
LENGTH out_line $200;
RETAIN out_line;
INFILE datalines TRUNCOVER;
INPUT @2 a_line $100.;
test = SUBSTR(a_line,1,7);
IF test = "MgrMsg:" THEN DO;
     out_line = a_line;
     INPUT @2 a_line $100.;
     test = SUBSTR(a_line,1,7);
     DO WHILE (test = ' ');
       out_line = CATX(' ',TRIM(out_line), TRIM(a_line));
       input @2 a_line $100. ;
test = SUBSTR(a_line,1,7);
        END;
        OUTPUT;

END;
ELSE DO;
out_line = a_line;
OUTPUT;
END;
RETURN;

DATALINES;
MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST
         *** FIRST LINE NEEDED *****
         *** MAY OR MAY NOT BE ANOTHER *****
APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed
MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO
MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO
MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/
         Jobno(43228) User(SCMSIS47) Host(SCMSIS
MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST
         User(SCPPSB38) Host(scppsb38)
APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed
MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/

         Jobno(59568) User(SCMSIS47) Host(SCMSIS
MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/
         Jobno(7800) User(SCMSIS47) Host(SCMSIS4
;
PROC PRINT;

**** Here is the output

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST
*** MAY OR MAY NOT BE ANOTHER *****
PPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed
MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO
MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/
MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST
PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed
MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/
MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/

Re: Info needed from input dataset wraps onto multiple lines.

a079011 — Mon, 19 Sep 2016 15:02:40 GMT

Thank you for the time you spent to respond to my question. Unfortunately this is not working as I expected. My input data looks like this.

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST

*FIRST LINE NEEDED*

*MAY OR MAY NOT BE ANOTHER*

APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO

MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO

MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/

Jobno(43228) User(SCMSIS47) Host(SCMSIS

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST

User(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/

Jobno(59568) User(SCMSIS47) Host(SCMSIS

MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/

Jobno(7800) User(SCMSIS47) Host(SCMSIS4

Output looks like this

*** MAY OR MAY NOT BE ANOTHER *****

APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO

MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO

Jobno(43228) User(SCMSIS47) Host(SCMSIS

User(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

Jobno(59568) User(SCMSIS47) Host(SCMSIS

Jobno(7800) User(SCMSIS47) Host(SCMSIS4

Desired output would be:

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST*FIRST LINE NEEDED**MAY OR MAY NOT BE ANOTHER*

APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed

MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO

MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO

MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/Jobno(43228) User(SCMSIS47) Host(SCMSIS

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN STUser(SCPPSB38) Host(scppsb38)

APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/Jobno(59568) User(SCMSIS47) Host(SCMSIS

MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/Jobno(7800) User(SCMSIS47) Host(SCMSIS4

Thank you again for your time 🙂

Re: Info needed from input dataset wraps onto multiple lines.

Shmuel — Mon, 19 Sep 2016 15:31:20 GMT

I have the feeling that small change will help:

On each INPUT line change from $100. to $char100.

Changing this informat will calculate (substring) TEST realy from pos 1.

Using $100 makes left justification and thus give fault TEST value.

Re: Info needed from input dataset wraps onto multiple lines.

a079011 — Mon, 19 Sep 2016 16:39:51 GMT

So close, it is only wokring on the first observation.

MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST * FIRST LINE NEEDED **MAY OR MAY NOT BE ANOTHER *

MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO

MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/ Jobno(43228) User(SCMSIS47) Host(SCMSIS

User(SCPPSB38) Host(scppsb38)

PPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed

MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/ Jobno(59568) User(SCMSIS47) Host(SCMSIS

Jobno(7800) User(SCMSIS47) Host(SCMSIS4

Re: Info needed from input dataset wraps onto multiple lines.

Shmuel — Mon, 19 Sep 2016 19:07:12 GMT

I made some changes to the program and tested it. Here is the code and the test data.

Pay attention: I had to add at end of datalines a new line to indicate the end of input lines (by TEST = 'EOF').

Without this indication, last output out_line won't be written.

Better check it again;

DATA want(keep=out_line);
LENGTH out_line $200;
RETAIN out_line ' ' flag_out 0;
drop flag_out;

INFILE datalines TRUNCOVER;
INPUT a_line $char100.;
test = SUBSTR(a_line,2,7);
if test = "MgrMsg:" THEN link process1; else
if test = " " THEN link process2; else
if test ne 'EOF' THEN link process3; else output;
RETURN;
*---------------------------*;
process1:
if flag_out = 1 then output;
out_line = a_line;
flag_out = 1;
RETURN;
*---------------------------*;
process2:
out_line = CATX(' ',TRIM(out_line), TRIM(a_line));
flag_out = 1;
RETURN;
*---------------------------*;
process3:
if flag_out = 1 then output;
out_line = a_line;
OUTPUT;
flag_out = 0;
RETURN;

DATALINES;
MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56910/MAIN ST
*** FIRST LINE NEEDED *****
*** MAY OR MAY NOT BE ANOTHER *****
APPLMGR: APPL DFUSCBM3.56910 JOB WFUSCBM3 Failed
MgrMsg: SCPPSB3L SCPPSB3L/SRVMON.2057/MAIN RESPO
MgrMsg: SCPPSB3M SCPPSB3M/SRVMON.2057/MAIN RESPO
MgrMsg: SCMSIS47 KPPSCDPP.QA029/DPPSC5@.1308642/
Jobno(43228) User(SCMSIS47) Host(SCMSIS
MgrMsg: SCPPSB38 WFUSCBM3/DFUSCBM3.56912/MAIN ST
User(SCPPSB38) Host(scppsb38)
APPLMGR: APPL DFUSCBM3.56912 JOB WFUSCBM3 Failed
MgrMsg: SCMSIS47 KPPSCDPP.QA014/DPPSC5@.1308642/
Jobno(59568) User(SCMSIS47) Host(SCMSIS
MgrMsg: SCMSIS47 KPPSCDPP.QA010/DPPSC5@.1308642/
Jobno(7800) User(SCMSIS47) Host(SCMSIS4
EOF
;

run;
PROC PRINT;

Re: Info needed from input dataset wraps onto multiple lines.

a079011 — Mon, 19 Sep 2016 19:54:01 GMT

This works perfectly! Thank you again for all the help.