topic Re: DATA merge Help Please in SAS Programming

DATA merge Help Please

SASHelpPlease — Thu, 12 Apr 2012 22:48:23 GMT

Hi there,

I'm having trouble with merging two data sets - one is a large file, bcarrier_line, with over 22 million rows. The other file is a DESY_SORT_KEY file with 344 rows (one column). I wish to merge the two files by DESY_SORT_KEY, so bcarrier_line file only outputs rows with desired DESY_SORT_KEY values.

I've ran this code with no errors, but the log output shows no row number change upon the merge (i.e. the larger file doesn't decrease in row size). From similar, but other, files I know the merge should significantly decrease the number of rows of the bcarrier_line data set.

CODE:

libname aaa 'M:\';
%let bcarrier = bcarrier_line;

*Already a SAS dataset;

DATA desy;
INFILE 'M:\DESY_SORT_KEY.csv';
INPUT DESY_SORT_KEY;
run;

PROC SORT data = aaa.&bcarrier;
by DESY_SORT_KEY;
run;

PROC SORT data=desy;
by DESY_SORT_KEY;
run;

DATA mergeaaa;
merge aaa.&bcarrier
desy;
by DESY_SORT_KEY;
run;

PROC EXPORT DATA = mergeaaa
OUTFILE='M:\bcarrier_line_aaa.csv';
run;

Any tips or advice is most appreciated!

Re: DATA merge Help Please

Reeza — Thu, 12 Apr 2012 23:31:02 GMT

Use SQL instead for a one step process.

PROC SQL;

Create table mergeaaa as

select * from aaa.bcarrier

where DESY_SORT_KEY in (select DESY_SORT_KEY from desy);

quit;

OR try the in option in your merge:

DATA mergeaaa;
merge aaa.&bcarrier (in=a)
desy(in=b);
by DESY_SORT_KEY;

if b;
run;

Re: DATA merge Help Please

SASHelpPlease — Fri, 13 Apr 2012 15:09:17 GMT

Hi Reeza,

Thanks - got it working!

For some reason I thought the "in" command with DATA MERGE was optional. As well, thanks for the related PROC SQL code.

Robin