Able to pipe query into stdin and get csv on stdout

We use commandline tools to transfer data from postgresql, mysql and sqlserver to (in our case) a postgresql server.

The basic idea is like this (in this case we get data from mysql and import it in postgresql:

# schema.table was created to have the same layout as whatever is getting out of the csv
echo "SELECT ... FROM ...; " | /bin/bash -o pipefail -c "\
MYSQL_PWD=password mysql --user=projecta --host=x.x.x.x --port=xxxx   --skip-column-names | \
PGOPTIONS='--client-min-messages=warning' PGTZ=Europe/Berlin \
psql --username=xxx --host=hostname --no-psqlrc --set ON_ERROR_STOP=on database --echo-queries \
--command=\"COPY schema.table FROM STDIN WITH NULL AS 'NULL'\"";

[we also run all our ETL via sending queries to stdin of a psql call, but to support that in mssql-cli would also mean “import csv on stdin” support to actually get the data into sql server]

For sqlserver, we currently use sqsh but supports at least the “SQL Script in, csv out” workflow. As sqsh is unmaintained (and buggy in versions available in ubuntu/debian), we would like to replace this by whatever finally supports at least exporting csv on stdout.

[sbrandtb]

This could be split into

• query from stdin (or file)
• csv stdout

The first one is keeping me from replacing microsoft’s crappy sqlcmd client with mssql-cli

[MrMeemus]

@JamesCass, thanks for the comment! Yes we are planning on updating the tool so we can target those specific scenarios regarding automation and scripting. A rough ETA would be some time March - April time frame.