CODEWITHSUNDEEP
pythonloopsdictionary
TomSanders
TomSanders

Reputation: 13

python: add one element to an existing dictionary from list loop

I need to export data from one database into another database. The dataobject needs to be mapped to a dictionary.

each row needs to be saved twice with different values from a list foobar=['foo','bar'].

foobar=['foo','bar']
data = []
for q in queryset:
    row = {"id": q.id,
           "created_at": q.created_at}
    for f in foobar:
        row['index'] = f
        data.append(row)
bulksave(data)

this doesn't give the desired result:

 print data
 [{'id': 1, 'created_at': '2017-01-01', 'index': 'bar'},
  {'id': 2, 'created_at': '2017-01-02', 'index': 'bar'}]

Where the desired output would have 'foo' and 'bar' as the index. How do I get this to work? Taking a different approach would be to loop differently:

for f in foobar:
    for q in queryset

this works, but takes twice the time because each element in the queryset will be evaluated twice.

Upvotes: 0

Views: 208

Answers (1)

tdelaney
tdelaney

Reputation: 77337

The problem is that inner loop. You keep updating the same dict and re-adding it to data. data ends up with multiple references to the same dict and the dict has been updated to the last value in foobar.

Just copy before adding. Also, a note on good question writing... you want a runnable example when possible. Here I mock queryset so the thing runs.

class Mock:
    def __init__(self, id, created_at):
        self.id = id
        self.created_at = created_at

queryset = [Mock(1, '2017-01-01'), Mock(2, '2017-01-02')]

foobar=['foo','bar']
data = []
for q in queryset:
    row = {"id": q.id,
           "created_at": q.created_at}
    for f in foobar:
        row['index'] = f
        data.append(row.copy())
bulksave(data)

UPDATE

If you really want to confuse the next person to look at your code, collapse it all into

import itertools

class Mock:
    def __init__(self, id, created_at):
        self.id = id
        self.created_at = created_at

queryset = [Mock(1, '2017-01-01'), Mock(2, '2017-01-02')]

foobar=['foo','bar']

bulkupdate(list(dict(zip(('id', 'created_at', 'index'), (q.id, q.created_at, index)))
    for q, index in itertools.product(queryset, foobar)))

(dgg32's now-deleted post got me thinking...)

Upvotes: 1

Related Questions