Check you logs. Teradata tools like FASTLOAD will eliminate duplicate records, but it should report that in the log.
You might check your SAS data for duplicate rows. Or it is possible that the variables in your TD table have transformed one or more variables and caused records to appear to be duplicates.
You can use PROC COMPARE , but it will mean that SAS will have to pull all 344K records out of TD to compare.
Might be easier to summarize on key variables (patient id, date, etc) and compare the summaries to see if there is a pattern to the loss.