The zipWithIndex() transformation appends (or ZIPs) the RDD with the element indices. This is very handy when wanting to remove the header row (first row) of a file.

Look at the following code snippet:

# View each row within RDD + the index 
# i.e. output is in form ([row], idx)
ac = airports.map(lambda c: (c[0], c[3]))
ac.zipWithIndex().take(5)

This will generate this result:

# Output
[((u'City', u'IATA'), 0),  
 ((u'Abbotsford', u'YXX'), 1),  
 ((u'Aberdeen', u'ABR'), 2),  
 ((u'Abilene', u'ABI'), 3),  
 ((u'Akron', u'CAK'), 4)]

To remove the header from your data, you can use the following code:

# Using zipWithIndex to skip header row
# - filter out row 0
# - extract only row info
(
    ac
    .zipWithIndex()
    .filter(lambda (row, idx): idx > 0)
    .map(lambda (row, idx): row)
    .take(5)
)

The preceding code will skip the header, as shown as follows:

# Output
[(u'Abbotsford', u'YXX'),  
 (u'Aberdeen', u'ABR'),  
 (u'Abilene', u'ABI'),  
 (u'Akron', u'CAK'),  
 (u'Alamosa', u'ALS')]