pandaq

pandaq: An easy pandas query-string builder

This library provides q method for easy querying of pandas.DataFrame. Internally, q generates the query string for pandas.DataFrame.query.

The goal of pandaq is to save time when querying.

Installation

pip install pandaq

Usage

pandaq provides two ways to select pandas.DataFrame rows by query string.

A. Generate a query-string

from pandaq import Q
import pandas as pd
df = pd.read_csv('https://raw.githubusercontent.com/datasciencedojo/datasets/master/titanic.csv')
qstr = Q().q(PassengerId=1)  # -> "PassengerId==1"
df.query(qstr)

B. Add q method to pandas.DataFrame

import pandaq.patch
import pandas as pd
df = pd.read_csv('https://raw.githubusercontent.com/datasciencedojo/datasets/master/titanic.csv')
df.q(PassengerId=1)

How to construct a query

Basics

Query Type	How It Works
q("str")	Equivalent to pd.DataFrame.query("str")
q(dict)	Works like q(**dict)
q(k=v)	Where the column named k is equal to v
q(k=[v1, ...])	Where the column named k is in [v1, ...]
q(k=(op1, v1, ...))	Where the column named k meets the condition k op1 v1 and ...
q(k="!str")	Where the column named k is NOT str (see Advance)
q(k="?str")	Where the column named k contains str (see Advance)
q(k="!?str")	Where the column named k does NOT contain str (see Advance)
q(k="/str")	Where the column named k match the regex str (see Advance)

Example

Type	Kind	pandaq.q	Equivalent to query-string
Number	Equal	q(a=1)	"a==1"
Number	Contain	q(a=[1, 2])	"a==1 or a==2"
Number	Inequality	q(a=(">", 1))	"a>1"
Number	Inequality	q(a=(">", 1, "<=", 3))	"a>1 and a<=3"
Bool	Equal	q(a=True)	"a==True"
Str	Full-Match	q(a="text")	'a=="text"'
Str	Not Full-Match	q(a="!text")	'a!="text"'
Str	Partial-Match	q(a="?text")	'a.str.contains("text", regex=False, na=False)'
Str	Not Partial-Match	q(a="!?text")	'not a.str.contains("text", regex=False, na=False)'
Str	Regex-Match	q(a="/textA|textB")	'a.str.contains("textA|textB", regex=True, na=False)'
Datetime	Equal	q(a=dt.date(1970,1,1))	'a=="1970-01-01"'
Datetime	Inequality	q(a=(">", dt.date(1970,1,1)))	'a>"1970-01-01"'

Combination

pandaq.q	Equivalent to query-string
q(a=1, b=1)	"(a==1 & b==1)"
q(a=1).q(b=1)	"a==1 & b==1"

Advanced Usage

To set the prefix individually, configure the following settings.

from pandaq import config

# Default settings
config.str_prefix = StringPrefix(
    neq="!",  # The prefix for not equal
    regex="/",  # The prefix for regular expressions
    partial="?",  # The prefix for partial match
    neq_partial="!?",  # The prefix for not equal partial match
)

License

pandaq is distributed under the terms of the MIT license.