A text file is always read with a DATA step, so you must write the code for it. Writing code can be automated through different means, e.g. CALL EXECUTE or writing code to a temporary file you later call with %INCLUDE.
To automate code writing, you need to have the necessary information for creating the dynamic parts of the code in data.
Take an example for a text-reading DATA step, and inspect it for the parts that need to be made dynamic. Then you wiill know what to put into a metadata dataset.
... View more
You can parse the output of the unzip command for the "XPORT" word and then extract the filename:
if index(_infile_,"XPORT")
then call symputx("fname",scan(_infile_,-1));
... View more
From the documentation, it seems that a physical filename must be used for the target file, so you cannot use a file reference created with the ZIP option there. This means you have to unzip first.
... View more
Do a Google search (Maxim 6) for „sas how to read csv file“ and you‘ll see that no further paper is needed. Reading data from text (flat) files has been around since SAS started, so it‘s among the best documented tasks.
... View more
@NewUsrStat wrote: Thank you very much for your explanation. I tried converting the Excel file into a CSV, but some variable names became "VAR*". I have to say that the variable names contain special characters.
This does not matter. In the DATA step you write, you will skip the header with FIRSTOBS=2 in the INFILE statement, and name the variables in the INPUT statement yourself.
... View more
The usual advice also applies here: save the spreadsheet to a text (csv) file and read that with a DATA step, which gives you full control over how the data is read.
... View more
Double DO method:
data want;
do until (last.var2);
set have;
by var2 notsorted;
if var1 = "XX" then flag = 1;
end;
do until (last.var2);
set have;
by var2 notsorted;
if flag then output;
end;
drop flag;
run;
... View more
The simplest method: create separate programs for the files and run them in batch mode in parallel, e.g. in UNIX:
sas program1.sas &
sas program2.sas &
sas program3.sas &
The & sends the command into background.
If you have a scheduler, it can handle this also.
... View more
It‘s possible, and there are multiple ways to do it. But you need to be careful so you do not saturate your resources (storage, network). Reading external data into datasets is always I/O bound.
... View more
As soon as it does what it is supposed to do, in reasonable time and without overloading the resources (CPU, storage, memory), it‘s good for release. But it should regularly be reviewed for possible improvements.
... View more
