Search on this forum as this question has been answered many times already. Here is a synopsis of best solutions in order or prefernce. 1) Fix it at the source if you can. That is have whoever gave your the file with embedded line breaks create version that does not have the embedded line breaks. 2) See if SAS can handle it anyway.  If you are lucky the lines are terminated with CR+LF characters and the embedded line breaks are using just LF (or just CR).  If so just adding the TERMSTR=CRLF option to your INFILE statement will let SAS read it without any extra work. 3) Run a data step to copy the original file to a new fixed file.  This is where you might try to use the COUNTW() trick to figure out when you have embedded line breaks and remove them. Another method is to count the quotes on the line and if a line break occurs when the number of quotes is odd then remove the line break.  Once you have a fixed version of the file with only one line per observation then just read it normally using data step. Or you could even let PROC IMPORT try to guess what the variable names and types are for you.     ... View more

First of all, I'm wondering what you expect from the new ID. Is there anything important that you cannot do with the existing ID?   As you say, these IDs are "coming from client". So, if there were any discussions about, say, data issues, you'd always have to refer to their IDs and not to yours. Also, updates of the data would contain the original type of IDs again and again and your conversion would have to work on these unknown future IDs as well.   If you really want to do it, you need a conversion which is guaranteed to be 1:1. I think a translation table, i.e. a dataset containing old and new IDs as its two columns, would be the safest way to handle this. You could easily check it for duplicates, amend it with future IDs, use it for the conversion (in either direction) by match-merging, possibly create a SAS (in)format for convenient in-line conversions etc.     ... View more

Thank you!!  I will try to tell him that.  I tried execute some simple commande on other tables and it worked fine.  Must just be the problem with that table. Thank you again. ... View more

Devon59 The SAS code on which your posting was based, looked like the style generated internally by PROC IMPORT. Although this works, it is not the easiest to support modify or explain. Thank you for posting some data. I know you already have a working solution, so this reply intends only to show how simple a solution can be achieved when you know your raw data layout.[pre]data loaded ; infile 'your csv file.csv' dsd firstobs= 2 ; length ID 8 Project $4 ; format DepositDate ReceiptDate datetime. amount comma12. ; informat DepositDate ReceiptDate YMDDTTM. ; input id -- amount ; run;[/PRE] Once you have defined or implied the columns of input with LENGTH and (IN)FORMAT statements, you can use the "--" variable list notation to simplify the input statement. Since your example data shows no time values, you may prefer the date-time format DTDATE9. I hope this example might help the next time you need to read a CSV file. PeterC ... View more