topic Re: Reading in a .csv file: a period added to a variable? in SAS Procedures

Reading in a .csv file: a period added to a variable?

ksio21008 — Tue, 09 Apr 2013 17:57:01 GMT

Dear All,

I'm trying to reading in a .csv file, but getting an error message on reading in one variable. The variable is a person's ID number. When I read in the file using an infile statement, I see a note saying that "Invalid data for this variable." In the log, I see a period added to this variable, which does not exist in the original .csv file. For instance, I see "84031563578" in the .csv file, but I see ".84031563578" in the SAS log instead. I don't know where this period came from. Can anyone suggest how I can fix this problem? Thank you!!!

SJK

Re: Reading in a .csv file: a period added to a variable?

ArtC — Tue, 09 Apr 2013 18:26:52 GMT

This could be an informat issue. Take a look at your INFORMAT for that field. Can we see your INPUT statement?

Re: Reading in a .csv file: a period added to a variable?

Ron_MacroMaven — Tue, 09 Apr 2013 21:21:46 GMT

"Id.Number" indicates that the column contains only digits.

This is a far-out suggestion:

Compare your id to

%Put SysMaxLong:&SysMaxLong;

Perhaps your id.numbers are greater than SAS can represent

which causes it to place the 'number' in a decimal format such as you are seeing.

Appropriately it ought to have a leading zero: 0.84031563578

next: 64-bit SAS

Ron Fehd id maven

Re: Reading in a .csv file: a period added to a variable?

ksio21008 — Tue, 09 Apr 2013 21:37:32 GMT

Thank you all! I fixed an informat statement, and now it works! I also assigned a proper number of places for the variable.

But I got a new problem. I have a series of data sets - 11 data sets that represent a data set from each month from Feb 2012 to Dec 2012.

I tried to read in all files at once by using the FILENAME statement, for example,

*FILENAME indat (

'C:\Users\sjk845\Desktop\data\feb2012.csv'

...

'C:\Users\sjk845\Desktop\data\dec2012.csv'

);

and refer indat in my INFILE statement.

DATA all;

INFILE indat DSD TRUNCOVER FIRSTOBS=2 TERMSTR=CR;

............

The problem is that it treats the header from files other than the first referred file (i.e., mar2012 - dec2012) as observation and gives me an error message. For example, it reads the first row of the March data as an observation, which is in fact the header.

Should I make 11 different INFILE statements for data for each month? Is there an optional syntax in the FILENAME statement that I can specify that the first row of each external files should be treated as header?

Thank you!

SJK