quarrel 0.6

quarrel makes it easy to query legacy rdbmses

querulous_quarrel

Named for a lovely groups of sparrows that happened to be flying by. A library that makes writing and executing queries a little easier for data scientists.

quarrel uses concentric and waddle and is proudly sponsored by the m&a team at cscgh.

installation

cd /path/to/repo
pip install quarrel

quick start

1. create a config file

    oracle:
      host: oracle.example.com
      user: scott
      password: tiger
      sid: xe
   

2. initialize concentric with the config file from (1)

    from concentric.managers import setup_concentric
   
    setup_concentric('/path/to/oracle.yml')
   

3. (optional) initialize the sql template directories

    from quarrel.settings import setup_quarrel
    setup_quarrel('/path/to/jinja2/sql/templates', '/path/to/jinja2/sql/queries') 
   

4. query the database

    from quarrel.raw import query
    results = query('oracle', 'select sysdate from dual')
   

raw results -- get raw results from the dbapi connection

quarrel.raw allows you to get the tuples as they were returned by the underlying dbapi connection. the header will be the cursor.description returned from the query, and the results will be the list of tuples returned by the query.

> from quarrel.raw import query
> header, results = query('oracle', 'select sysdate d from dual')
> print(header[0][0])
D
> print(results)
[(datetime.datetime(2022, 6, 26, 15, 10, 59),)]

cooked results -- get results as a list of dicts

quarrel.cooked allows you to get a list of dicts which is slightly easier to understand and work with but can be substantially slower as python will construct a dict per row that is returned. Each key of the dict will be the lower-cased column name specified in the query.

> from quarrel.cooked import query
> results = query('oracle', 'select sysdate d from dual')
> print(results)
[{'d': datetime.datetime(2022, 6, 26, 15, 14, 12)}]

sqlalchemy results

quarrel.sqlalchemy allows you to get a list of dicts as well. However, it uses sqlalchemy under the hood. This can be useful when you need sql alchemy's connection pooling features.

> from quarrel.sqlalchemy import query
> results = query('oracle', 'select sysdate d from dual')
> print(results)
[{'d': datetime.datetime(2022, 6, 26, 15, 14, 12)}]

pandas results

quarrel.pandas allows you to get a dataframe using either the dbapi connection or a sqlalchemy connection.

> from quarrel.pandas import query
> results = query('oracle', 'select sysdate d from dual')
> print(results)
                    d
0 2022-06-26 15:20:34

> from quarrel.pandas import query_alchemy
> results = query_alchemy('oracle', 'select sysdate d from dual')
> print(results)
                    d
0 2022-06-26 15:20:34

In order to use pandas, make sure you install pandas support using the pandas extra

pip install quarrel[pandas]