Joining with wild card key, PROC SQL?

kjohnsonm — Tue, 29 Mar 2016 17:14:33 GMT

Hello all,

I have two lists one of 6100 files a fully qualified windows dir\path\filename.ext with random punctuation like: &,()'

and another larger list 81K obs. The 6100 records will each and everyone have a match in the larger list however the larger list has none of those punctuation marks any longer but does have a place holder underscore. "_" if any punctuation was taken out for example:

D:\path\file&name.ext D:\path\file_name.ext

D:\path\path's\file&name.ext D:\path\path_s\file_name.ext

E:\path (my)\pat&h\new, file.SAV E:\path _my_\pat_h\new_ file.SAV

etc.

6100 81K

Does anyone know how to crosswalk "left join" these two lists with wild cards and might have the time to toss me a slow pitch solfball?

/*for a clear list of what has been compressed out*/
compress(Path_File, "',&()", "")

These 6100 files are the last of 240K files I need to research for metadata, however I did not have the skill to make SAS read these files with the problematic punctuation still in the file paths/names. PS the path file names can be upto 260 char long, all other data is derived from the path, file name and data type so no fields were given as examples. TIA. -KJ

Re: Joining with wild card key, PROC SQL?

PGStats — Mon, 28 Mar 2016 22:30:22 GMT

Use function translate()

a inner join b on a.newPath = translate(b.oldPath,"____", "&,()")

Re: Joining with wild card key, PROC SQL?

kjohnsonm — Tue, 29 Mar 2016 17:12:39 GMT

proc sql;
create table testing as
Select a.Path_File1 as keya,
	   b.Path_File1 as keyb
from small_data_set a
inner join large_data_set b
on a.Path_File1 = translate(b.Path_File1,"_____", "&',()")
;
quit;

The data sets are exactly 4 files off and I hand checked them earler and removed their obs because of side issues, thanks for the help.

topic Joining with wild card key, PROC SQL? in SAS Data Science

Joining with wild card key, PROC SQL?

Re: Joining with wild card key, PROC SQL?

Re: Joining with wild card key, PROC SQL?