Backup of David's Livejournal

My list of addresses has made its way from a physical address book, to a Palm Pilot, to Microsoft Outlook, to Google Contacts to the iPhone Contacts app. Along the way, each of the transitions has played fast and loose with the mappings of the individual fields.

Nowadays, the normal way for a Windows user to export contacts from an iPhone is to sync between the iPhone and Microsoft Outlook, then export from Outlook to a CSV file. I hate having to go through that middleman.

I prefer to take stronger ownership of my own data, and have settled on Google Contacts as the primary place for my data. There are a few reasons. I really like the open "group" (tagging) feature of contacts. I like that they have a public API for accessing and manipulating contacts. But most important is the ability to easily export and import contacts. As Joel Spolsky noticed ten years ago, a good way to get me to try a service, is to make it easy for me to change my mind and leave the service.

By default only the contacts group named "My Contacts" will sync with the iPhone when you set up an Exchange sync between the two end points. That suits my purposes just fine.

I've taken some time to groom my contacts list for Google Contacts. Here are some notes from that experience:

When you export contacts, use the Google CSV format. Your contacts will be exported to a UTF-16 file, and all the special characters you use will be retained. If you choose Outlook CSV format, then the file generated will be 8-bit regardless of the characters used in your contact list, and characters that don't map to 8-bit characters will be changed to question marks. So for 安室奈美恵's sake, choose Google CSV format.

Most people edit their CSV files in a spreadsheet editor. That's fine, but I don't trust my own eyes and hands to get everything right, so I prefer to do batch editing programmatically.

If you want to do some batch processing of the CSV file in Python, here are some snippets. These snippets have been pared down to the essentials, and don't represent good coding practices.

To read in the Google CSV:

# unicode_csv_reader and UnicodeWriter are provided
# in the documentation from the cvs module.

f = unicode_csv_reader(codecs.open( 'google.csv', 'r', 'utf-16' ))

headings = f.next()
col = {}
for i in range(len(headings)):
    col[headings[i]] = i

rows = []
for row in f:
    rows.append(row)

If you want to print out a list of contacts sorted by last name:

rows.sort(key=operator.itemgetter(col['Family Name']))

Beverly Howard suggests that contacts with no "Name" field but only a "Company" field won't sync with Outlook. You could check for that:

if row[col['Organization 1 - Name']] and not row[col['Name']]:
    # Ensure these get synced across all devices, some don't!

Finally, after you've made your batch changes, you can write them back to a UTF-16 file like so:

out_file = open( 'google_out.csv', 'wb' )
google_out = UnicodeWriter( out_file, encoding='utf-16' )
google_out.writerow( headings )
google_out.writerows( rows )

It's been a long time coming, but I'm glad I've got more of my data in a place where it's easy for me to get it out and manipulate it any way I like.