Python

Storage space for my Python code that may or may not be in production.

Syntax cheat-sheet

Convert ISO 8601 datetime string into a datetime object

import dateutil.parser
new_date = dateutil.parser.parse(string_date)

Create a complete file path using a directory and a file name

from pathlib import Path
file_path = Path(directory_var).joinpath(file_name_var)

Delete a key from a dictionary

dict.pop("name_of_key", None)

Delete duplicates from a list (does not retain the original order)

list2 = list(set(list1))

Delete None values from a dictionary

new_dict = {k: v for k, v in dict.items() if v is not None}

Execute a Python script in ipython

exec(open("python_script.py").read())

Get the character represented by an ordinal value

ord_char = chr(ord(orig_char))

Get current date

import datetime
current_date = datetime.datetime.today()

Get difference in days between two datetime objects

import datetime
date_diff = dt_object2 - dt_object1
date_diff.days

Get the longest string in a list of strings

result = max(string_list, key=len)

Get the name of a file from a string containing the full path

from pathlib import Path
in_file = Path(full_path_var).name

Get the name of a file, excluding extension, from a string containing the full path

from pathlib import Path
in_file_no_ext = Path(full_path_var).stem

Get the ordinal value of a character

char_ord = ord(orig_char)

Pandas - append data frames

df1 = df1.append(df2, ignore_index=True, sort=False)

Pandas - append dictionary to data frame

df = df.append(dictionary, ignore_index=True)

Pandas - check that a column contains a string

(df["column_name"].notnull()) &
(df["column_name"].str.contains("string"))

Pandas - check that a column does not contain a string

(df["column_name"].notnull()) &
(~df["column_name"].str.contains("string", na=False))

Pandas - convert a column to a list

column_list = df["column"].tolist()

Pandas - convert a dataframe to a list of dictionaries

dict_list = df.to_dict("records")

Pandas - convert values in a list column to string

df["column_name"] = ["".join(i) if isinstance(i, list) else i for i in df["column_name"]]

Pandas - convert values in a string column to integer

df["column_name"] = pd.to_numeric(df["column_name"], errors="coerce").astype(pd.Int64Dtype())

Pandas - count the number of occurences of each unique column value

unique_values = df.groupby("column_name")["column_name"].count()

Pandas - count the total number of unique column values

total_unique = df["column_name"].nunique()

Pandas - create new column based on current column values

df.loc[df["current_column"] == "CurrentColumnValue", "new_column"] = "NewColumnValue"

Pandas - drop columns

df.drop(["column1", "column2", etc.], axis=1, inplace=True)

Pandas - drop duplicate rows (keep first row and drop all others)

df.drop_duplicates(keep="first", inplace=True)

Pandas - drop duplicate rows based on columns (keep first row and drop all others)

df.drop_duplicates(["column1", "column2", "column3", etc.], keep="first", inplace=True)

Pandas - drop NaN

df.dropna(axis=0, subset=["ColumnName"], inplace=True)

Pandas - get the first X columns in a dataframe

df1 = df.iloc[:, :X]

Pandas - get only alphabetic characters from a string

df["new_column"] = df["original_column"].str.replace("\d+", "", regex=True)

Pandas - get only numeric characters from a string

df["new_column"] = df["original_column"].str.extract("(\d+)", expand=False)

Pandas - initialize empty data frame with column names

df = pd.DataFrame(columns=["column1", "column2", "column3", etc.])

Pandas - keep specified columns

df = df[["column1", "column2", etc.]]

Pandas - merge data frames (outer) with indicator

new_df = pd.merge(df1, df2, how="outer",
                  left_on="df1_column", right_on="df2_column",
                  indicator=True)

Pandas - read_csv specifying utf-8 (get rid of special characters in the first column name)

df = pd.read_csv(filehandle, encoding="utf-8")

Pandas - read_csv with tabs as delimiters

df = pd.read_csv(filehandle, sep="\t")

Pandas - rename columns

df = df.rename(columns={"OldColumnName1": "NewColumnName1", "OldColumnName2": "NewColumnName2", etc.})

Pandas - replace Greek letter mu with a "u"

df["column"] = df["column"].str.replace('\xb5', 'u')

Pandas - replace NaN with empty string

df.fillna("", inplace=True)

Pandas - replace NaN with None

df2 = df1.where(df1.notnull(), None)

Pandas - split a column and take the first/last elements

df["new_column"] = df["original_column"].str.split("split character").str[0]
df["new_column"] = df["original_column"].str.split("split character").str[-1]

Pandas - strip unwanted characters from a column

df["column_name"] = df["column_name"].str.strip("unwanted characters")

Pandas - write to an output file

df.to_csv(filehandle, index=False)
df.to_csv(filehandle, sep="\t", index=False)

Pandas - write to an output file in DOS without creating blank lines

filehandle = open("filename", "w", newline="\n")
df.to_csv(filehandle, index=False)

Pandas - write to standard output

import sys
df.to_csv(sys.stdout, index=False)

Reload a module in ipython

from importlib import reload
reload(module)

Remove all non-numeric characters in a string

import re
new_string = re.sub("[^\d]", "", old_string)

Remove duplicates from a list

new_list = list(set(old_list))

Rename a key in a dictionary

dictionary[new_key] = dictionary.pop(old_key)

Run a script from interactive Python (V3)

exec(open("python script").read())

Set the random seed to the current datetime

import random
from datetime import datetime
random.seed(datetime.now().timestamp())

Synapse login

import synapseclient
syn = synapseclient.Synapse()
syn.login(silent=True)

Synapse walk a Synapse folder structure to get file SynIDs

Synapse login (see above)
import synapseutils
syn_contents = synapseutils.walk(syn, folder_syn_id)
for __, __, filelist in syn_contents:
    if len(filelist) > 0:
        for __, syn_id in filelist: