01-23-2015 11:46 PM
I previously had a problem when I used proc import to import a csv file into SAS EG (two of the numeric variables were detected for errors for some rows). But when I imported the csv file manually, it was fine. I was suggested by Xia Keshan that I could use guessingrows=32767 statement. When I tried the proc import with this statement, I didn't have any errors like before. However, just noticed that now all the numeric variables including sas date variables all became character variables. Any suggesion on resolving this issue? Thanks.
01-24-2015 02:57 AM
"just noticed that now all the numeric variables including sas date variables all became character variables"
That is beacuse these numeric variabls and date variables contains some characters like TAB ......
You need to change them on your own, and no doubt you get some error in log, and use ?? or compress() to avoid the error info.
01-25-2015 02:53 PM
Thanks Xia keshan. Yes, you are right. Now I imported them on my own.
Just out of curiosity. Did you mean those variables might have ..... or actual word TAB? I guess .... right? I do see that those two variables have decimal points.
01-26-2015 01:37 AM
I don't know. But they do have some characters I believe.
If you use input() to convert them into numeric variable, You are going to get ERROR message in log window , I think.
12-01-2016 12:03 PM
Can you give some small example on how can we read .csv files using datalines and without guessing rows and proc import?
It will be of great help.