Condition

Project Description

This project provides a user friendly way to construct a condition object. The condition object can be passed around and later used in various contexts, such as querying pandas Dataframe, filtering pyarrow partitions or generating where clauses in SQL. It takes care of formating and syntax for you. It also supports extensibility to new usage contexts.

A Quick Example

>>> from condition import *
>>> import pandas as pd
>>> fl = FieldList('A B C date value'.split())
>>> cond = ((fl.date >= pd.to_datetime('20000101')) 
...             & (fl.date <= pd.to_datetime('20010131'))
...             & (fl.A == 'a1 a3'.split()) \
...             & (fl.B == 'b1')) \
...             | (fl.C != 'c3 c5'.split())
>>> print(cond) # show as sql where condition

        (
                (date >= '2000-01-01 00:00:00'
                and date <= '2001-01-31 00:00:00'
                and A in ('a1','a3')
                and B = 'b1')
        or C not in ('c3','c5'))

>>> cond.to_df_query()
"(((date >= '2000-01-01 00:00:00')&(date <= '2001-01-31 00:00:00')&(A in ('a1','a3'))&(B == 'b1'))|(C not in ('c3','c5')))"
>>> cond.to_pyarrow_filter()
[[('date', '>=', Timestamp('2000-01-01 00:00:00')),
  ('date', '<=', Timestamp('2001-01-31 00:00:00')),
  ('A', 'in', {'a1', 'a3'}),
  ('B', '=', 'b1')],
 [('C', 'not in', {'c3', 'c5'})]]

SQL Generation

>>> from condition.sql import render_sql
>>> sql = """
...     select *
...     from my_table
...     where {{where_condition}}
... """
>>> print(render_sql(sql, cond))

    select *
    from my_table
    where 
        (
                (date >= '2000-01-01 00:00:00'
                and date <= '2001-01-31 00:00:00'
                and A in ('a1','a3')
                and B = 'b1')
        or C not in ('c3','c5'))

Please see Usage Examples for detailed examples.

Installation

This project is distributed via pip. To get started:

pip install condition

To install jinja2 package used by condition.sql, do the following

pip install "condition[sql]"

To install all packages for development, do the following

pip install "condition[dev]"