Map Latitude and Longitude for each data point

I have a lot of timeseries data for 30 stations and I need to map the latitude and longitude to each data point so I can plot it on a map. The latitude and longitude data is in another .csv that needs to be maped over based on the stations ID. I know I can write a very long for loop, but I was hoping to find a more efficient wat to look it up in the other .csv and fill it in for each data point.

Here is a summary of the data with just 3 stations for a few hours of data:

date    precip_mm   temp_C  ID
11/2/2006 11:00 47.7    3.8 301
11/2/2006 12:00 47.7    4.4 301
11/2/2006 13:00 48.9    4.4 301
11/2/2006 14:00 50.2    3.8 301
11/2/2006 15:00 50.2    2.9 301
11/2/2006 16:00 50.2    2.7 301
11/2/2006 17:00 50.2    2.8 301
11/2/2006 18:00 50.2    2.6 301
11/2/2006 19:00 50.2    2.5 301
11/2/2006 11:00 50.2    2.2 321
11/2/2006 12:00 50.2    1.7 321
11/2/2006 13:00 50.2    1.5 321
11/2/2006 14:00 50.2    1.2 321
11/2/2006 15:00 50.2    1.1 321
11/2/2006 16:00 50.2    1.3 321
11/2/2006 17:00 50.2    1.3 321
11/2/2006 18:00 50.2    1.3 321
11/2/2006 19:00 50.2    1.2 321
11/2/2006 11:00 50.2    1.4 391
11/2/2006 12:00 50.2    1.2 391
11/2/2006 13:00 50.2    1.1 391
11/2/2006 14:00 50.2    1.1 391
11/2/2006 15:00 50.2    1.6 391
11/2/2006 16:00 50.2    2.1 391
11/2/2006 17:00 50.2    2.7 391
11/2/2006 18:00 51.4    2.1 391
11/2/2006 19:00 51.4    0.9 391

Here is the metadata:

Name       ID            Latitude   Longitude
Adin Mtn    "301    "   41.23583    -120.79192
Bear Creek  "321    "   41.83384    -115.45278
Cedar Pass  "391    "   41.58233    -120.3025

I have tried to map with the following code and all of the forums I have seen are matching one item to one item with melt or merge but I need to fill in the lat and long for many years of data. When i run this like of code, the column in filled with nan and does not give me an error. Help please!

df['latitude'] = df['ID'].map(metadata.set_index('ID')['Latitude'])

>Solution :

Most likely the reason you are getting NaN’s is because ID is stored as a string in your metadata but as an integer in your timeseries data. You need to convert that column to a proper numerical column for the merge to match properly.

meta['ID'] = meta['ID'].astype('int')
df.merge(meta, on = 'ID')