Create a gist now

Instantly share code, notes, and snippets.

What would you like to do?
Elasticsearch Python bulk index API example
>>> import itertools
>>> import string
>>> from elasticsearch import Elasticsearch,helpers
es = Elasticsearch()
>>> # k is a generator expression that produces
... # a series of dictionaries containing test data.
... # The test data are just letter permutations
... # created with itertools.permutations.
... #
... # We then reference k as the iterator that's
... # consumed by the elasticsearch.helpers.bulk method.
>>> k = ({'_type':'foo', '_index':'test','letters':''.join(letters)}
... for letters in itertools.permutations(string.letters,2))
>>> # calling k.next() shows examples
... # (while consuming the generator, of course)
>>> # each dict contains a doc type, index, and data (at minimum)
>>> k.next()
{'_type': 'foo', 'letters': 'ab', '_index': 'test'}
>>> k.next()
{'_type': 'foo', 'letters': 'ac', '_index': 'test'}
>>> # create our test index
>>> es.indices.create('test')
{u'acknowledged': True}
>>> helpers.bulk(es,k)
(2650, [])
>>> # check to make sure we got what we expected...
>>> es.count(index='test')
{u'count': 2650, u'_shards': {u'successful': 1, u'failed': 0, u'total': 1}}

zyxwu commented Apr 3, 2016

but when I index documents one by one, everything's fine
res = es.index(index=INDEX, doc_type=DOC_TYPE, id=ind, body=JS_message)

k2xl commented Mar 22, 2017

Would this work for a python list of JSON documents?

@zyxwu Butt it's hell slow.

I have a use case where I am updating/adding documents using the bulk API call and just after I fire the bulk call I check the ES count using the count API. Problem is there is some delay that happens after firing the bulk call which results in a delay in reflecting the correct count. How do I make sure the correct count reflects after I fire the bulk call?

@tusharkale197 ElasticSearch is near real time database not exact real time database and hence once you index a document changes will be reflected within index.refresh_interval time period default value of which is 1s you can change it accordingly and also if you want you can manually refresh index using refresh api

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment