Python For Data Science Cheat Sheet Excel Spreadsheets Pickled Files

>>> file = 'urbanpop.xlsx' >>> import pickle

Importing Data >>> data = pd.ExcelFile(file) >>> with open('pickled_fruit.pkl', 'rb') as file:
pickled_data = pickle.load(file)
>>> df_sheet2 = data.parse('1960-1966',
Learn Python for data science Interactively at www.DataCamp.com skiprows=[0],
names=['Country',
'AAM: War(2002)'])
>>> df_sheet1 = data.parse(0, HDF5 Files
parse_cols=[0],
Importing Data in Python skiprows=[0], >>> import h5py
>>> filename = 'H-H1_LOSC_4_v1-815411200-4096.hdf5'
names=['Country'])
Most of the time, you’ll use either NumPy or pandas to import >>> data = h5py.File(filename, 'r')
your data: To access the sheet names, use the sheet_names attribute:
>>> import numpy as np >>> data.sheet_names
>>> import pandas as pd Matlab Files
Help SAS Files >>> import scipy.io
>>> filename = 'workspace.mat'
>>> from sas7bdat import SAS7BDAT >>> mat = scipy.io.loadmat(filename)
>>> np.info(np.ndarray.dtype)
>>> help(pd.read_csv) >>> with SAS7BDAT('urbanpop.sas7bdat') as file:
df_sas = file.to_data_frame()

Text Files Exploring Dictionaries

Stata Files Accessing Elements with Functions
Plain Text Files >>> data = pd.read_stata('urbanpop.dta') >>> print(mat.keys()) Print dictionary keys
>>> filename = 'huck_finn.txt' >>> for key in data.keys(): Print dictionary keys
>>> file = open(filename, mode='r') Open the file for reading print(key)
>>> text = file.read() Read a file’s contents Relational Databases meta
quality
>>> print(file.closed) Check whether file is closed
>>> from sqlalchemy import create_engine strain
>>> file.close() Close file
>>> print(text) >>> engine = create_engine('sqlite://Northwind.sqlite') >>> pickled_data.values() Return dictionary values
>>> print(mat.items()) Returns items in list format of (key, value)
Use the table_names() method to fetch a list of table names: tuple pairs
Using the context manager with
>>> with open('huck_finn.txt', 'r') as file:
>>> table_names = engine.table_names() Accessing Data Items with Keys
print(file.readline()) Read a single line
print(file.readline()) Querying Relational Databases >>> for key in data ['meta'].keys() Explore the HDF5 structure
print(file.readline()) print(key)
>>> con = engine.connect() Description
>>> rs = con.execute("SELECT * FROM Orders") DescriptionURL
Table Data: Flat Files >>> df = pd.DataFrame(rs.fetchall()) Detector
>>> df.columns = rs.keys() Duration
GPSstart
Importing Flat Files with numpy >>> con.close()
Observatory
Files with one data type Using the context manager with Type
UTCstart
>>> filename = ‘mnist.txt’ >>> with engine.connect() as con:
>>> print(data['meta']['Description'].value) Retrieve the value for a key
>>> data = np.loadtxt(filename, rs = con.execute("SELECT OrderID FROM Orders")
delimiter=',', String used to separate values df = pd.DataFrame(rs.fetchmany(size=5))
df.columns = rs.keys()
skiprows=2,
usecols=[0,2],
Skip the first 2 lines
Read the 1st and 3rd column
Navigating Your FileSystem
dtype=str) The type of the resulting array Querying relational databases with pandas
Magic Commands
Files with mixed data types >>> df = pd.read_sql_query("SELECT * FROM Orders", engine)
>>> filename = 'titanic.csv' !ls List directory contents of files and directories
>>> data = np.genfromtxt(filename, %cd .. Change current working directory
%pwd Return the current working directory path
delimiter=',',
names=True, Look for column header
Exploring Your Data
dtype=None)
NumPy Arrays os Library
>>> data_array = np.recfromcsv(filename) >>> data_array.dtype Data type of array elements >>> import os
>>> data_array.shape Array dimensions >>> path = "/usr/tmp"
The default dtype of the np.recfromcsv() function is None. >>> wd = os.getcwd() Store the name of current directory in a string
>>> len(data_array) Length of array
>>> os.listdir(wd) Output contents of the directory in a list
Importing Flat Files with pandas >>> os.chdir(path) Change current working directory
pandas DataFrames >>> os.rename("test1.txt", Rename a file
>>> filename = 'winequality-red.csv' "test2.txt")
>>> data = pd.read_csv(filename, >>> df.head() Return first DataFrame rows
nrows=5, >>> os.remove("test1.txt") Delete an existing file
Number of rows of file to read >>> df.tail() Return last DataFrame rows >>> os.mkdir("newdir") Create a new directory
header=None, Row number to use as col names >>> df.index Describe index
sep='\t', Delimiter to use >>> df.columns Describe DataFrame columns
comment='#', Character to split comments >>> df.info() Info on DataFrame
na_values=[""]) String to recognize as NA/NaN >>> data_array = data.values Convert a DataFrame to an a NumPy array DataCamp
Learn R for Data Science Interactively