Pandas - using PostCoder to lookup Latitude & Longitude in each row, then return Postcode in new columns

Question

I want read two columns (Latitude & longitude) from my dataframe df1 in pandas and create a new column zipcode and add zipcode at each row in the dataframe.

I think this webpage is useful: https://postcodes.readthedocs.io/en/latest/

df1 = df[['Col1',' Col2', 'Col3','Col4', 'Col5', 'Latitude', 'Longitude']]

for row in df1[7]:
    # Try to,
    try:
    # get lat long and find the post code
        postcodes.get_nearest(lat, lng)

    # But if you get an error
    except:
        # error

# Create a new columns post code in df1
df1['postcode'] = zipcode

score 0 · Accepted Answer · answered Apr 01 '19 at 02:34

You have to use apply to create new column based on other data of dataframe.

def getPostcode(row):
    try:
        row['postcode']=postcodes.get_nearest(row['Latitude'], row['Longitude'])
    except:
        print('Error for data {0}'.format(row))
    return row

Then add this line to main code after init df1:

df1.apply(getPostcode,axis=1).

score 0 · Answer 2 · answered Apr 01 '19 at 02:40

0

You can try:

df1['postcode'] = df1.apply(
    lambda x: postcodes.get_nearest(x['Latitude'], x['Longitude']),
    axis=1
)

You can imagine that the apply function loops every row or column of the dataframe executing a function (in this case, the lambda function).
It will loop rows when the axis option is 1, and will loop columns when axis options is 0 (default).
This lambda function receives a row as x, then it sends the 'Latitude' and 'Longitude' values to .get_nearest.

Depending on the size of your dataframe, it may take a while though.
I've tested the postcodes here and it didn't worked, but if this lib is working for you, this code should do fine.

answered Apr 01 '19 at 02:40

edinho

406
2
6

You can not use inline try-except in lambda block. – Serenity Apr 01 '19 at 03:18
I wasn't intending to use try-except. But if it is the case, it can define a function instead of using lambda. – edinho Apr 01 '19 at 03:21
Thanks @edinho. .get_nearest is not working for me too. Is there any other way beside using the PostCodes library. – Ashish Apr 03 '19 at 17:15
I don't know. But I would start by looking at open street map project and google apis. – edinho Apr 04 '19 at 01:08

Pandas - using PostCoder to lookup Latitude & Longitude in each row, then return Postcode in new columns

2 Answers2