bruno-uy

import pandas as pd

# Create sample pandas.Series to calculate frequency
s = pd.Series(pd.date_range("2021", freq="18H", periods=1000))
# Calculate frequency grouping by month
# If you want another period check following link:
# https://pandas.pydata.org/docs/user_guide/timeseries.html#timeseries-offset-aliases
freq_count = s.dt.to_period("M").value_counts(sort=False)

bruno-uy

import pandas as pd

# The function shows `str` for dates, but could be datetime.date objects as well
def n_date_intervals(start_date: str, end_date: str, intervals_count: int) -> pd.Series:
    return pd.Series(pd.date_range(start_date, end_date, periods=intervals_count))

# Change this variables as needed
start_date = "2018-01-01"
end_date = "2021-03-24"
intervals_count = 5

bruno-uy

-- You can count the rows to be equal to columns that you're checking
SELECT *
FROM information_schema.columns
WHERE table_schema = '<table_schema>'
	AND table_name = '<table_name>'
	AND column_name IN ('<column1>', '<column2>');

bruno-uy

from functools import partial

def print_n_dict_elem(n, dictionary, elem):
    print(n * dictionary[elem])

# Create a new function that print elem 3 times
print_3_dict_elem = partial(print_n_dict_elem, n=3, dictionary={"a": "A", "b": "B"})
print_3_dict_elem(elem="a")

# Update dictionary passed as an argument

bruno-uy

# Remember to install more_itertools first
# pip install more-itertools
from more_itertools import grouper

def do_something_with_iterable(iterable):
    pass

n = 3
chunks = grouper('a'*10, n)
for c in chunks:

bruno-uy

import pandas as pd

df = pd.DataFrame({"A": [0, 1, 0], "B": ["a", "b", "c"], "C": [1, 2, 3]})
df.set_index(["A", "C"], drop=False, inplace=True)
result = {level: df.xs(level).to_dict("index") for level in df.index.levels[0]}
print(result[0])

bruno-uy

import pandas as pd

utc_now_pd = pd.Timestamp.utcnow()
# The function round(freq="D") is magic behind the round up
# Use replace(tzinfo=None) to remove timezone information
utc_now_ceil = utc_now_pd.round(freq="D").to_pydatetime().replace(tzinfo=None)
# Convert to ISO format
utc_now_str = utc_now_ceil.strftime("%Y-%m-%dT%H:%M:%S")

bruno-uy

import pandas as pd
df = pd.concat([pd.read_csv(f, compression="gzip") for f in os.listdir() if f.endswith(".gz")], ignore_index=True)

bruno-uy

import pandas as pd
import numpy as np

df = pd.DataFrame({"A": [1, 2, 3], "B": [1.2, np.NaN, 3.4]})
result = (
    df
    .replace([np.nan], [None], regex=False)
    .to_dict(orient="records")
)

bruno-uy

-- Problem: you don't see all the schemas when querying PG_TABLE_DEF
-- Solution:
-- 1. First check if the schema you're trying to query is on the search path
show search_path;
-- 2. Add the missing one(s) to the search path (imagine the result was only public and you're missing data_warehouse and matching)
set search_path to '$user', public, data_warehouse, matching;  -- No matter which is your user, use '$user'