@Andrew-M  While the SAS/ACCESS engines in all their flavors try to avoid truncation due to hard limit of 32767 in Base SAS data set (*.sas7bdat), specially when dealing with extra large character columns, this could be a double-edge thing!! it could lead to over/under provisioning!! Add to it, Byte vs. Character representation -- I had that issue with Snowflake and utf-8 encoding! That's why I would try to profile my source table(s) either before (if possible) or after ingestion, and change the structure of my SAS data set to make it as lean as possible, using a custom developed SAS macro program.  You can  find it here on GitHub sasCensusApi/code/macros/etl/etl_shrinkmydata.sas at main · ahmedanadwc/sasCensusApi   Hope this helps, Ahmed ... View more