I haven't used the Merge transformation that much, but it's quite straight forward - but it's' linked to the syntax of SQL MERGE statement, so you need to know what you want to do. Why do you have to use the Merge transformation? if you are updating a SCD Type 2 table, there's a transformation for that... ... View more

Thanks a lot. This works perfectly! ... View more

Just confirming you want to delete a sheet in a workbook, not the whole workbook? If it's the whole workbook, then you can use FDELETE. ... View more

Hello Patrick   I have attached my 'Have' dataset (from the SQL step), as well as the cases and controls datasets. Thanks for the help. ... View more

Suggest: Provide a bit better description of your data. You can do that with Proc Freq and NLEVELS to report on the number of levels of each variable. Example: ods select nlevels; proc freq data=sashelp.class nlevels; run; The ODS select means that only the Nlevels table is generated which should not run out of memory but will give us an idea just how much stuff you are dealing with. "Blank" is not a SAS value you will see. Depending on which procedure/approach you are dealing with you can use the MISSING option to get a count.   You can run into memory issues when generated large tables because the table with all the border information and such are stored in memory before writing out the results. If you have the default HTML destination and are sending the data to another ods destination that can seriously increase the memory needed. So you might have better results by closing the HTML destination and writing to only one destination like ODS RTF of PDF.   One basic syntax would look like: proc tabulate data=yourSetNameHere; class _character_/missing; table _character_, n pctn; run; but missing values will be at the top of each as the default order for character values is alphabetic and blanks come before any other printable character.   ... View more

Thank you so much @AMSAS ! Your code worked. I just had to do a small tweak to the following original code: /* Need to know what month the data_date is in */ data work.have2 ; set work.have ; mm=month(data_date) ; dow=putn(data_date,"downame3.") ; run ; Since I have multiple years of data for each 'seccode', I had change the 'mm' variable to reflect both month and year: /* Need to know what month the data_date is in */ data work.have2 ; set work.have ; mm=100*year(data_date)+ month(data_date); dow=putn(data_date,"downame3.") ; run ; That seems to work. I appreciate all your help with this! ... View more

Thanks much, that helped .../svi-datahub/admin/lists?name=region_districts_sub_lev did the trick. ... View more

If you have a metadata-enabled SAS installation on Linux which means you have a working metadata tier, then the easiest way to run batch jobs in my opinion is to use the Schedule Manager functionality in SAS Management Console. If you aren't familiar with SMC, it is a PC-installed tool that allows you to connect to your SAS server and run or schedule batch jobs via a GUI - you avoid using command lines completely.   Schedule Manager automatically builds and runs a SAS batch command line behind the scenes and puts all of the batch logs in one folder, nicely time and date stamped so they are all in one place and in time order. ... View more

NO. I do not think that is possible. SAS only have string and number as datatypes. ... View more