Data Handling using Pandas – I

pandas Series and DataFrames: making them, selecting with loc and iloc, filtering, changing and saving them.

Key ideas

A Series is a 1-D labelled array (labels 0, 1, 2, … by default); a DataFrame is a 2-D table of labelled rows and columns.
loc selects by label, slicing INCLUDING the end label; iloc selects by position from 0, end left out.
df.loc['S2', 'Marks'] or df.iloc[1, 3] is one value; df.loc['S2'] is a row labelled by column names.
head(n), tail(n): the first or last n rows (5 by default); df[df['Marks'] > 50] keeps the matching rows.
Arithmetic on a Series works element by element and keeps the labels; pd.Series(5, index=[…]) repeats 5.
df['Total'] = df['A'] + df['B'] adds a column; df.drop(label) drops a row, df.drop(col, axis=1) a column.
shape is (rows, columns) and size is rows × columns; pd.read_csv() reads a CSV file, df.to_csv() writes one.

Remember

Labels a to f	s.loc['b':'d'] → b, c, d; s.iloc[1:3] → b, c
Some columns	df[['Name', 'Maths']] (two pairs of brackets)
shape (r, c), size r × c	add a column: c + 1; drop a row: r − 1; drop a column: c − 1

Watch out

Leaving out the end label in a loc slice, or keeping the end in an iloc slice.
Forgetting axis=1 when dropping a column: drop() removes a row by default.

Worked example

Series s:
a | 38
b | 96
c | 75
d | 29
e | 61
f | 56
Write the output of each statement, and say why they differ.
(a) print(s.loc['b':'f'])
(b) print(s.iloc[1:5])

Working: loc slices by label (end included); iloc by position (end left out)

Answer: (a)
b | 96
c | 75
d | 29
e | 61
f | 56
(b)
b | 96
c | 75
d | 29
e | 61
loc includes the end label f; iloc stops before position 5.