dbpath
Automation of database scripts: turning results to jq to simply scripting for support and diagnostics
Motivation
It is mediumly challenging to execute sql scripts with command line scripts. One reason is that sql is extremely verbose
This project introduces dbpath
that allows us to specify queries in a more concise way, and these queries
can then be executed from the command line, or a script or a program.
Current state:
We have a command line tool dbpath
that 'knows' a little about the database structure. This knowledge is
in a file called dbpath.json
. This file is in the current directory or a parent.
The tool just does simple queries that joins tables together. It is intended for support people that need to walk tables.
Often these people often have a notepad full of common queries that they copy and paste into the database tool. This
tool is intended to subliment that, and make it easy to do the same thing from the command line.
dbpath ? # lists the tables
dbpath d? # lists the tables that start with d
dbpath driver 123 # lists the driver with id 123
dbpath driver --name phil # lists the driver with name phil
dbpath driver.audit 123 # lists the audit records for driver 123 (the records for driver are joined to the audit records)
dbpath driver.mission.audit 123 # lists the audit records for the missions that driver 123 has been on
dbpath driver.mission.audit 123 -date '2023-6-3' # lists the audit records for the missions that driver 123 has been on for the given date
The sql queries are derived from a knowledge of the database that is in the dbpath.json
file. Currently,
this has to be populated manually, but I expect to auto generate it from the database schema.
Features
-s or --sql or --fullSql
Show the sql instead of executing it. FullSql includes the paging sql (which can be 'noisy')
dbpath driver 123 -s
# select T0.*
# from drivertable T0 where T0.driverId=123
dbpath driver 123 --fullSql
# select T0.*
# from drivertable T0 where T0.driverId=123
# LIMIT 15 OFFSET 0
-t or --trace
Execute the command one step at a time, showing the sql and the results
dbpath driver.mission.audit 123 -ts
-w or --where
Add a where clause to the query. Typically you would do a -s
first to find the alias name
dbpath driver.mission.audit -w 'audit.id = 123'
dbpath driver -w 'T0.name = "phil"' # Note the need for the quotes. This is because the shell is parsing the command line
-c or --count
Show the count of the results instead of the results
dbpath driver --count
# count
# 2
dbpath driver --count --notitles
# 2
--distinct
Show the distinct values of the results
--notitles
Don't show the titles on results. This can help a lot when making scripts that use the output
dbpath driver --notitles
# 1 phil
# 2 joe
--json
Show the results as json. This is useful for piping to jq
dbpath driver --json
# [
# {
# "driverid": 1,
# "name": "phil"
# },
# {
# "driverid": 2,
# "name": "joe"
# }
# ]
##--onelinejson
dbpath driver --onelinejson
# {"driverid":1,"name":"phil"}
# {"driverid":2,"name":"joe"}
Environments
Typically, we have multiple environments. The tool can be configured to use different databases for different environments:
this is in the dbpath.json
file. The default environment is dev
(and in current state that's the only one usable)
The environment gives the database type and the connection details
dbpath envs
# Current environment is dev
# Environment Type Host Port Database UserName
# dev postgres localhost 5432 postgres phil
# test postgres localhost 5432 postgres phil
Current environment
This defaults to 'dev'.
It can be changed to another legal value by
dbpath env test
Checking the environments are accessible
dbpath status
#Environment Type Host Port Database UserName Up
#dev postgres localhost 5432 postgres phil true
#test postgres test.example.com 5432 postgres phil true
Secrets
The username and password don't need to be specified if they are in environment variables:
- DB_AUTO__USERNAME provides the username. For example
export DB_AUTO_DEV_USERNAME=phil
- DB_AUTO__PASSWORD provides the password. For example
export DB_AUTO_DEV_PASSWORD=phil