Detecting merge steps in pseudocode mappings

kalbo — Fri, 06 Aug 2021 09:30:21 GMT

Hi,

I'm trying to write some code that will scan pseudo sas code stored in a mapping sheet which merges or concatenates data sets.

In example 1 I want to see if there is "data" followed by "merge" and "run".

The lines "ANY TEXT..." just mean that any other code or text may be present in those lines.

Thanks.

****example 1****

ANY TEXT DKASJLFLDLSAJFKLDFJD;
DATA NEW;
MERGE DSONE(IN=A) DSTWO(IN=B);
RUN;
ANY TEXT DKASJLFLDLSAJFKLDFJD;

In example 2 I want to see if there is "data" followed by "set" followed by at least 2 data sets, and want to ignore the (in=) parts.

****example 2****

ANY TEXT DKASJLFLDLSAJFKLDFJD;
DATA NEW;
SET DSONE(IN=X) DSTWO(IN=Y) DSTHREE;
RUN;
ANY TEXT DKASJLFLDLSAJFKLDFJD;

In example 3 I want to see if there is "proc sql" followed by "join" and "quit"

****example 3****

ANY TEXT DKASJLFLDLSAJFKLDFJD;
PROC SQL;
select customers.city, sales.totals
from sales natural inner join customers
where country='United States';
QUIT;
ANY TEXT DKASJLFLDLSAJFKLDFJD;

Re: Detecting merge steps in pseudocode mappings

ballardw — Fri, 06 Aug 2021 15:11:32 GMT

Followed how closely for cases one and two?

A data step may any number of statements before a Set or Merge depending on coding style. For example you might have someone assign a bunch of lengths, formats or labels for some reason.

There are also some techniques with multiple SET statements that are used to combine/interleave more than one data set.

Does every single data step end with a Run? If not you may be opening can worms for logic.

And for a truly memorable question, what does the expected output from the process look like? I have no idea what you may want/expect the output to look like.

topic Re: Detecting merge steps in pseudocode mappings in SAS Programming

Detecting merge steps in pseudocode mappings

Re: Detecting merge steps in pseudocode mappings