遍历字典和字符串列表

Question

I have a list of dictionaries converted from JSON but few keys seem to be Unicode which giving me trouble accessing the keys of the dictionary. The list looks like this:

d = [{'location': u'',
      'partner_id': '648746',
      'partner_is_CSP': 'Y',
      'partner_name': 'D4T4 SOLUTIONS PLC',
      'partner_programs_tiers': [{'program_name': 'Cloud Service Provider',
        'tier_name': 'Gold'}],
      'partner_type': 'Direct Reseller; Service Provider',
      'sort_value': '60',
      'url_to_logo': u'',
      'url_to_website': 'https://www.d4t4solutions.com/'},
     {'location': {'address_type': 'H',
       'city': 'Tirane',
       'country': 'ALBANIA',
       'geo_latitude': '41.348335',
       'geo_longitude': '19.79865',
       'phone': u'',
       'point_of_contact': u'',
       'state': u'',
       'street_1': 'RR. E DURRESIT PALL. M.C.INERTE KATI 1 LAPRAKE',
       'street_2': u'',
       'street_3': u'',
       'zip': '1023'},
      'partner_id': '649341',
      'partner_is_CSP': 'N',
      'partner_name': 'COMMUNICATION PROGRESS',
      'partner_programs_tiers': '[]',
      'partner_type': 'Distribution VAR',
      'sort_value': '0',
      'url_to_logo': u'',
      'url_to_website': 'www.commprog.com'}]

Now, I want to do something like this:

l = [i["location"].get("street_1",None) for i in d]

But I'm getting the following error:

AttributeError: 'Unicode' object has no attribute 'get'

How can I work my way around that Unicode? Thanks a lot for your help.

PS list d contains more dictionaries than shown here and it contains more than just one Unicode. When I iter over the dictionaries, I would like to have None value for the location key with an empty Unicode value is encountered.

Answer 1

You could use this (rather unreadable) one-liner:

>>> [r['location'].get('street_1', None) if isinstance(r['location'], dict) else (r['location'] or None) for r in d]
[None, 'RR. E DURRESIT PALL. M.C.INERTE KATI 1 LAPRAKE']

It's probably better to go with a full for-loop:

>>> l = []
>>> for r in d:
...     loc = r['location']
...     if isinstance(loc, dict):
...         l.append(loc.get('street_1', None))
...     else:
...         l.append(loc or None)
...
>>> l
[None, 'RR. E DURRESIT PALL. M.C.INERTE KATI 1 LAPRAKE']
>>>

Essentially, use isinstance to check if you are working with a dict or not. If we are, use .get , if we aren't, append the value. I use loc or None which will evalute to None if loc is not truthy, which u"" happens to be not truthy.

The alternative is a EAFP approach:

>>> for r in d:
...     loc = r['location']
...     try:
...         l.append(loc.get('street_1', None))
...     except AttributeError:
...         l.append(loc or None)
...
>>> l
[None, 'RR. E DURRESIT PALL. M.C.INERTE KATI 1 LAPRAKE']

Whether or not it is more efficient to go with that or a LBYL approach depends on the nature of the data. If the "exception" is not truly exceptional, ie it occurs frequently, then a LBYL approach will actually be faster, even if EAFP is considered Pythonic.

Answer 2

Just modifying your attempt a little, using an empty dict as default.

>>> [(i['location'] or {}).get('street_1') for i in d]
[None, 'RR. E DURRESIT PALL. M.C.INERTE KATI 1 LAPRAKE']

Answer 3

As simple way would be:

for i in d:
    location = i['location']
    if location:
        print(location.get('street_1', 'n/a'))  # or whatever you want to do...