08-14-2014 05:04 PM
Its a huge csv will file ith more number of variables .If I use infile I need to get all the formats correct to get the file to read .
08-14-2014 05:11 PM
Read the first 10 rows as 100 strings to review content (easier in columns).
Maybe increase that 10 if the first10 dont clarify informats.
Perhaps the supplier of a large set of data (in a csv file) could provide the data definition that defines what the informats should be.
08-14-2014 06:16 PM
Use proc import on the data with a large value for guessing rows. Look in the log for the code,
copy and modify as needed such as preferred date formats, add labels, change variables that maybe should be character that get imported as numeric such as Id numbers, zip codes, phone numbers.
I go through this at least once a month with new data sources.