Pandas - Manipulating data row-wise

import pandas as pd

iris = pd.read_csv("archive.ics.uci.edu/ml/machine-learning-dat..")

df = iris.copy()

df.columns = ['sl', 'sw', 'pl', 'pw', 'flower_type']

1) deleting a particular row

df.drop(0, inplace = True)

df.drop(3, inplace = True)

df.head()

It deletes the row with the label 3. By default, an extra column from 0 to n - 1 labeled column is added, which gets mistaken to be index values. In reality, it is just a label. Hence, if try to run the same above code again, it will throw an error as the row with label = 3 will get deleted in the first run.

If we don't use the inplace = True argument, all the changes made will be in the copy of df file not on the original df file.

2) see all the labels

print(df.index)

Int64Index([  1,   2,   4,   5,   6,   7,   8,   9,  10,  11,
            ...
            139, 140, 141, 142, 143, 144, 145, 146, 147, 148],
           dtype='int64', length=147)

3) deleting particular row based on position or actual index value

df.drop(df.index[0], inplace = True)

df.head()

4) deleting multiple rows at the same time

df.drop(df.index[[0, 1]], inplace = True)

df.head()

5) running a particular condition

df.sl > 5

5      False
6      False
7      False
8      False
9       True
       ...  
144     True
145     True
146     True
147     True
148     True
Name: sl, Length: 144, dtype: bool

This way returns in a true or false format

/ Better representation method

df[df.sl > 5]

116 rows × 5 columns

This method returns only those rows which satisfies the given condtion.

df[df.flower_type == 'Iris-setosa']

Generate more detailed information

df[df.flower_type == 'Iris-setosa'].describe()

6) checking a particular row

print(df.iloc[0]) # position based

print(df.loc[5]) # label based

sl                     4.6
sw                     3.4
pl                     1.4
pw                     0.3
flower_type    Iris-setosa

Name: 5, dtype: object
sl                     4.6
sw                     3.4
pl                     1.4
pw                     0.3
flower_type    Iris-setosa
Name: 5, dtype: object

7) adding a row

df.loc[0] = [1, 2, 3, 4, 'Iris-sertosa']

df.tail()

Adds a row to the last with the label name = 0 and provided data.