I use them for importing from files to tables. Run some cleanup and validation, before moving the data to its final destination. If pg crashes halfway, I still

I use them for importing from files to tables. Run some cleanup and validation, before moving the data to its final destination. If pg crashes halfway, I still have the file imported and restart from zero. A second use is importing httpd combined logs and doing some analysis when finding out which chains of http calls caused some kinds of behaviour. When done, the tables get deleted. It allows easy ad hock queries, correlating with monitoring tables, indexing... I used to write some python scripts, but pg did better than expected, and this way of working stuck somehow.