PROC IMPORT variable

[jen123]

Below is my code.

In file_a, there are 99 columnswith over 100K rows.

There is one column in this file that has either numeric or alpha data (not alpha-numeric) for each entry.  The format for this column in DATAROW=2 is numeric.  When I run this file, SAS thinks the column is numeric.  If it comes to a cell with ALPHA, it returns "." for missing.

How can I fix the code - without having to list EVERY INFORMAT, FORMAT and INPUT for all 99 variables?

Thanks!

********************************

PROC IMPORT OUT=LIBNAME.file_b

DATAFILE = 'file_a'

DBMS=CSV replace;

GETNAME=YES;

DATAROW=2;

RUN;

[Reeza]

I think it might be DBTYPE option, see the documentation...

SAS/ACCESS(R) 9.2 Interface to PC Files: Reference, Second Edition

[Doc_Duke · Accepted solution]

You also might be able to address this using the GUESSINGROWS statement in PROC IMPORT.

[jen123]

I didn't know how to use DBTYPE with PROC IMPORT.

I tried GUESSINGROWS with PROC IMPORT.  The results look like they match the input file - # of rows in dataset matches .csv file.  However, I got an error:

NOTE: Invalid data for Glob_Function_Id in line 94102 839-842.

Errors detected in submitted DATA step. Examine log.

108956 rows created in libname.file_b from file_a.csv.

ERROR: Import unsuccessful. See SAS Log for details.

I got many of the NOTE in the log.  I checked the line and everything looks good.  Any thoughts as to the error?  What do the numbers 839-842 mean?

Thanks!!!

[Doc_Duke]

839-842 are the columns for the variable that SAS had trouble reading.

[jen123]

I thought they refer to columns, but my file has 99 columns (variables).  So that confused me.

Anyway, I checked the entire row and everything matched with the original file.  So I am not sure what the error is.