Samuel Oranyeli samukweku

samukweku / pd_nested_grp.py

Last active October 12, 2025 06:34

pandas nested groupby limitations

data = pd.DataFrame({'id':[1,1,1,2,2,2],'sales':[4,1,2,7,6,7],'views':[3,1,2,8,6,7]})

samukweku / conditional_join.py

Last active November 30, 2022 20:10

Code for pydata global conditional_join talk

	# data is from DuckDB's github repo
	# https://github.com/duckdb/duckdb/tree/master/benchmark/micro/join

	import pandas as pd; import numpy as np; import janitor as jn; import duckdb
	query = """SELECT SETSEED(0.8675309);
	CREATE TABLE events AS (
	SELECT *,
	"start" + INTERVAL (CASE WHEN random() < 0.1 THEN 120 ELSE (5 + round(random() * 50, 0)::BIGINT) END) MINUTE
	AS "end"
	FROM (

samukweku / wide_to_long_datatable.py

Last active April 16, 2022 14:23

Idea for wide_to_long implementation in datatable

	from datatable import dt, f
	from typing import Pattern, NamedTuple, Union
	from collections import Counter, defaultdict
	from itertools import compress, chain
	import re
	import numpy as np

	class measure(NamedTuple):
	"""reshape either with a separator or a regular expression."""
	column_names:Union[str, list]

samukweku / pandas_nth.md

Last active August 2, 2021 08:51

Pandas Chaining - Second most common value

import pandas as pd

df = pd.DataFrame({'x': [1, 1, 1, 2, 2, 3],

samukweku / pandas_hype.md

Last active November 15, 2021 06:15

Solutions to some challenges on `Dont believe the python hype...`

import pandas as pd
import numpy as np
import janitor