On Mon, 20 Feb 2023 15:24:23 -0800 Adrian Klaver wrote: >On 2/20/23 11:36, pf@xxxxxxxxxxx wrote: >> On Mon, 20 Feb 2023 11:06:34 -0800 Adrian Klaver wrote: >> >>> On 2/20/23 10:27, pf@xxxxxxxxxxx wrote: >>>> [Still a newbie; but learning fast...] >>>> >>>> Hi, > >> >> Notwithstanding the man page, my take is that the DROP DATABASE statement >> needs to be eliminated at pg_dump creation by pgAdmin4. Taking this to >> that mailing list. > >It just dawned on me you might be doing all of this through the >pgAdmin4 GUI. Sorry for any confusion... I get it now... A team member uses pgAdmin4 to load separate table(s) into his DB; then creates dump files (one per table) of those _individual_ tables which are uploaded to me. I maintain a complete set of tables in my DB. pgAdmin4 is never used here; the restore is done with a simple bash script which inserts the dbname and dumpfile name into the command. Those restore tables should only be created with DROP DATABASE _off_. All that should happen with my DB is to add these tables (99.9% of the time, they are totally new to me). A DROP DATABASE from the one-table per dumpfile creator is UNwanted. Looks like I really did dodge a bullet... >Again the default is to not include those options. Glad to know pgAdmin4 has those switches. Thanks again!! >> >> Thanks Tom & Adrian! >> >>>> Was my 134 table[1] myname DB saved because it was open? >> >> Tom: Yup. >> >>>> If the dump file >>>> contains the above statements, how can I be absolutely certain I won't >>>> lose the DB? >> >> Tom: >> Reading the manual is advisable. --create --clean specifies exactly >> that the target database is to be dropped and recreated. >> >> regards, tom lane >> >>>> I'm obviously quite paranoid now... >>> >>> You will lose the database if you do as the docs specify for -C: >>> >>> " >>> -C >>> >>> ... >>> >>> When this option is used, the database named with -d is used only to >>> issue the initial DROP DATABASE and CREATE DATABASE commands. All data >>> is restored into the database name that appears in the archive. >>> " >>> >>> >>> It will then be recreated with whatever information is in "dumpfile". If >>> that is the same data or new data you want then you are fine. Otherwise >>> you will need to be more specific about what you are trying to achieve. >>> >>> >>>> >>>> [1] 3 types of tables: ~40%=8.5M rows; ~40%=33M rows; ~20%=varying sizes >>>> >>>> Thanks, >>>> Pierre >>>> >>>> >>> >> >> >
