Irina Truong

Same answer is here: https://github.com/dbcli/pgcli/issues/1191#issuecomment-663907983 it's not one of the commands pgcli supports. PRs are always welcome.

Thank you for filing the issue @DataMower! 1. You said in our previous conversation that you installed pgcli with conda. How are you running conda? Are you using Windows Subsystem...

How did you install pgcli? Was it with `pip`?

The fix is now merged to the main branch. Because it's not released yet, you could try it out by uninstalling pgcli from Homebrew, and installing it from github: ```...

This works with a file that contains plain sql commands in pgcli==2.1.0: ``` Server: PostgreSQL 10.6 Version: 2.1.0 Chat: https://gitter.im/dbcli/pgcli Mail: https://groups.google.com/forum/#!forum/pgcli Home: http://pgcli.com postgres@/tmp:postgres> \i 1.sql +-------+ | boo...

Hi @vanviegen , you need this change in `test_pgexecute` to fix the failing test: ```diff diff --git a/tests/test_pgexecute.py b/tests/test_pgexecute.py index 7fef4658..882c4845 100644 --- a/tests/test_pgexecute.py +++ b/tests/test_pgexecute.py @@ -527,6 +527,12 @@...

@laixintao I like the idea of `history_file` being linked to `alias_dsn`, and the possibility of sharing the file for databases that are located on different hosts but have the same...

Hah. I forgot that we have the same name already. You're right, it may be confusing. We better name the section differently. Maybe `history_file_mapping`, or something like that.

> Another approach I just thought of would be to enable appending -o optionName=value to CLI params, similarily to ssh and other commands @quezak This sounds like a good solution...

This still happens, but the problem does not look exactly the same. It only happens when writing with `pyarrow` and reading with `fastparquet`. And it looks like we're sometimes reading...