

I don't think it is optimized for memory usage any where near as well as an sde view. There is no way to get the tool to behave like a Keep All Records join, where every unmatched record displays Null values, so any unmatched features in your tables would cause the unmatched features to disappear in the result. With criteria it only keeps matching records which reduces it to about the number of records within the smaller table. If I had not specified any criteria for how the fields of the table relate to each other, the tool would have cross-multiplied every record in the FC and the table (700,000 * 36,000 = 25,200,000,000 combinations). It took about 7 minutes to process after I indexed the criteria fields on both the FC and the table (without indexed fields I killed it after 30 minutes). I have used this tool with the 700K features you are talking about to see a join to Building permits (only about 36K records). I would definitley stick with views using sde command line for performance on an FC of this size. Is this the expected behavior of the Make Table Query or perhaps I am misusing this tool, and should stay with sde command line, which takes a few seconds to process this much data. The process ran for 23 hours before I killed it because it sucked up about 23 GBs of memory in my Local Setting Temp directory.

I recently tried to use the Make Query Table, thinking that its kinda like the Spatial Views in sde command line, but in ArcMap, and I was using a layer with over 700,000 parcels, with multiple tables, which each have around a million records.
