@PGStats    Just so I'm clear...The guidance is to convert the variables in each individual dataset before joining, correct? ... View more

@Tom   This explanation is very helpful. Thank you.   After playing around w/ these ideas, I noticed that the lengths and formats for many of these variable conflicted, or there was no format assigned, as you mentioned. I also noticed that when the LENGTH and the FORMAT conflict, the LENGTH attribute "wins" in the dataset.    <You fix this by adding a format statement that lists variables but does not have a format specified. This will remove the format from those variables.>   In your example you put "_character_" in the FORMAT statement. What is the effect of this? Will it simply strip the format of all character variables?   Final question, if I included an INFORMAT statement in a DATA step like this, where would it go relative to the LENGTH, SET and FORMAT statements? After LENGTH and before SET? ... View more

@FreelanceReinh Sorry. I should have tried this first. Given my struggles lately, I assumed that it couldn't be that easy...   It works, kinda. A new PROC CONTENTS says the length has been changed, however the values are still truncated. Obviously I need to do some more investigating.    The values are not truncated in any of the datasets that go into the final dataset...Hmmm.   Btw, I should have made it clear that I am joining/stacking the 5 merged datasets into one final dataset (i.e. Not merging). I'm going to make the change in the original post. ... View more

@Astounding I've never used it. Would it be too much to ask for some example code?    Btw, I should have made it clear that was joining/stacking the 5 merged datasets into one final dataset (i.e. Not merging)... ... View more

@Reeza    <That's what GuessingRows specifies>   I understand. I was referring to a PROC IMPORT w/o using the GUESSINGROWS statement.   <the default value is 20>   Meaning the first 20 observations (assuming no GUESSINGROWS statement)?     ... View more

@ballardw Thanks for sharing your approach here. I think I'm going to adopt this approach going forward.   Quick question: Do you know how may observations SAS reads during a PROC IMPORT before it makes its determination about the attributes of a particular variable?  I'm thinking specifically about the length of character variables, but I'm interested in the answer for informats and formats also. Thanks! ... View more

Maybe "dictate" is too strong a word. Wouldn't it "trick" SAS into setting the variable type as character, and possibly use a relevant informat when reading-in the data? ... View more

<Out of interest, why would you want to use proc import to import CSV data?>   This question of GUESSINGRO WS arises out of a broader question. Using PROC IMPORT v. a DATA step to read-in your data. As I understand it, and I am by now means an expert here, the risk of using PROC IMPORT is that SAS "guesses" incorrectly on one or more variables, and this causes problems down the line. The risk of using a DATA step is that it may be a lot of work if you have many variables for which to set attributes. My dataset has 500 variables...   The thought behind the GUESSINGRO WS approach was that this might be a middle ground approach. Also, @Reeza offered, what sounds like, a good approach. Use PROC IMPORT and then just copy the syntax from the log and change the attritbutes on the variables that you want to change.    I am open to any other ideas or suggestions!   Thanks for your thoughts and help here. ... View more

@ballardw <Line 2 looks like this: YO,,,,,,,,,,,,,,,,SIXDIG,SIXDIG,,,,,>   This was intentional. It was my attempt to dictate that the variables w/ these character values (e.g. YO, SIXDIG) be read-in as character variables, despite the fact that their subsequent values are purely numeric.   <Using guessingrows you have to indicate at least the first datarow. >   I may not understand how to implement this GUESSINGROWS approach =;-)...I was trying to indicate the second row - the row w/ YO and SIXDIG - as the row that SAS should use to determine the data type.   <Your next to last row of data has invalid values for what are time values in the other rows. 101 is not going to be read as time when the rest are 12:00 AM type (when using proc import).>   Will the times be read in as times?! I thought that SAS would simply create character variables for these variables given that there is no AM/PM associated w/ the time in one observation. Also, in my larger dataset, there are some instances where there is no space between the last digit and the AM or PM (e.g. 12:00PM v. 12:00 PM).    Couldn't this problem also be addressed w/ the GUESSINGROWS approach, assuming I implemented it correctly? If I put a time like, 12:00 PM, in the GUESSINGROWS row, wouldn't that dictate the variable type (Character) and maybe the informat?   Thanks again for your help! ... View more

@Tom I uploaded an .xlsx file b/c it wouldn't let me upload a .csv file. I subsequently learned that I could simply change the extension to .txt. Sorry for the confusion. I uploaded a new file w/ a .txt extension. ... View more

  @ballardw @Tom Thanks for your help!   I'm getting an error w/ the following code. Is it possible this statement doesn't work in SAS University Edition?   PROC IMPORT DATAFILE="/folders/myfolders/...guessing_rows.csv" OUT=want DBMS=csv replace ; GUESSINGROWS=1; GETNAMES=yes; DATAROW=2; RUN; LOG: NOTE: Unable to open parameter catalog: SASUSER.PARMS.PARMS.SLIST in update mode. Temporary parameter values will be saved to WORK.PARMS.PARMS.SLIST. Number of names found is greater than number of variables found. Number of names found is greater than number of variables found. ERROR: Import unsuccessful. See SAS Log for details. NOTE: The SAS System stopped processing this step because of errors.   The data is uploaded as a .xlsx file, but I used it as a .csv. ... View more

@Vince28_Statcan   <name=upcase(substr(name,1,1))||lowcase(substr(name ,2,0))>   What do the "1,1" and the "2,0" represent in this syntax?    Thanks! ... View more