It looks like the database you are connecting to uses the UTF-8 character set, which has characters that are not part of the Latin1 character set that your SAS session uses. When EG encounters one of these, it throws the error. You might want to review this thread:
... View more
The default separator (at least in the German locale) is the semicolon when excel writes to a csv file. One has to make sure that the columns in the spreadsheet have a format assigned that corresponds to the informat one will use in the data step. Then the data will arrive healthy in the SAS table.
... View more
Most of the options the afore posters have given (libname excel, proc import) miss an important fact: they are only available on Windows, but not on UNIX. Export the data from excel as .csv, transfer the file to UNIX, and read it there. That way you have the additional advantage of controlling the interface (the .csv) with a simple text editor.
... View more
It is not unexpected behaviour, just see the SAS documentation for the input statement: SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition Go to "Input Styles - Column Input". Since it can do what the code told it to do without problems, SAS sees no need for NOTE, WARNING or ERROR messages.
... View more
Since "1" or "2" is not a format, but a value, and follows a variable name, it is interpreted by the input statement as a column. Therefore you read the contents of columns 2 and 1, respectively, which happen to be 2 and 7. An "informat" without a dot is not an informat, it is ALWAYS something else. The only time you write a format/informat name without a dot is when specifying it in proc format at creation time. Edited a typo
... View more
#2 is the best option, hands down. Everything else is crushing peanuts with an elephant's hoof, IMO. If you want to do it inside SAS, #1 is preferred. A simple "data xxx; set yyy; run;" will not be slower than #3, because it will be simply limited by I/O bandwidth. Hint: Copying inside the same library will usually be slower than copying to another one, because you use the same HW for reading and writing. Different libs _may_ be located on different disks, giving you more aggregate bandwidth.
... View more
If there's a different SAS installation per project, I can see this happening. But AFAIK the licenses will keep being done per machine, and priced according to number of cores / computing power.
... View more
Use the COUNTW function with a ',' delimiter to determine the number of occurences in a line. Then run a do loop for 1 to (result of COUNTW), use the SCAN function with delimiter ',' to get the nth value group, split that with another use of the SCAN function with default delimiter ' ', and output. KEEP only the wanted variables.
... View more
