how to do bulk indexing to elasticsearch from python

Question

I have nearly 10K json documents and i want to push all this documents to elasticsearch by using elasticsearch bulk api from python. I went through some docs but didn't get any solutions.

result=es.bulk(index="index1", doc_type="index123", body=jsonvalue)
helpers.bulk(es,doc) 

i tried both but no result,i am getting this error

elasticsearch.exceptions.RequestError: TransportError(400, u'illegal_argument_exception', u'Malformed action/metadata line [1], expected START_OBJECT or END_OBJECT but found [VALUE_STRING]')

please help me

Solution

I prefer using the bulk method present in helpers module for bulk indexing. Try the following:

from elasticsearch import helpers
res = helpers.bulk(es, jsonvalue, chunk_size=1000, request_timeout=200)

Your jsonvalue needs to follow a particular format. It needs to be a list of the 10K json documents with each document having the following fields:

doc = {
    '_index': 'your-index',
    '_type': 'your-type',
    '_id': 'your-id',
    'field_1': 'value_1',
    ...
}

So your final jsonvalue would look something like this:

jsonvalue = [
    {
    '_index': 'your-index',
    '_type': 'your-type',
    '_id': 'your-id',
    'field_1': 'value_1',
    ...
},
    {
    '_index': 'your-index',
    '_type': 'your-type',
    '_id': 'your-id',
    'field_1': 'value_2',
    ...
},
    {
    '_index': 'your-index',
    '_type': 'your-type',
    '_id': 'your-id',
    'field_1': 'value_3',
    ...
}
]

Comments

[RCP]

can't i write my json documents inside "_source": { }

[Jai Sharma]

_source convention is the way elasticsearch stores the data internally and returns it in this way. You should pass your json document at the first level in the dictionary.

[RCP]

sorry i am not getting you, please tell me where should i use 'doc' in helpers.bulk(), and where should i pass my json documnets?

[Jai Sharma]

Edited my answer.

[RCP]

yeah my json value is in above format , where should i pass that value and where should i pass 'doc' in helpers.bulk()?

[Jai Sharma]

helpers.bulk(es, jsonvalue, chunk_size=1000, request_timeout=200). The 2nd argument should be the list of all the documents, as mentioned in the answer.

[RCP]

so now in this helpers.bulk(es, jsonvalue, chunk_size=1000, request_timeout=200) which argument is holding "index name,doc_type" and all

[Jai Sharma]

Index name and doc type are now part of each document, _index and _type hold the values for index & doc type respectively. Please read the documentation elasticsearch-py.readthedocs.io/en/master/… here

[RCP]

Tried for sample data jsonvalue = { '_index': 'your-index', '_type': 'your-type', '_id': 'your-id', 'field_1':{"priority": "CRITICAL"} } helpers.bulk(es, jsonvalue, chunk_size=1000, request_timeout=200) but getting error elasticsearch.exceptions.RequestError: TransportError(400, u'action_request_validation_exception', u'Validation Failed: 1: index is missing;2: type is missing;3: index is missing;4: type is missing;5: index is missing;6: type is missing;7: index is missing;8)

[Jai Sharma]

Let us continue this discussion in chat.

[Ansuman]

from 2020 => '_index': 'your-index', '_type': 'your-type', '_id': 'your-id', 'field_1':{"priority": "CRITICAL"} } helpers.bulk(es, jsonvalue, chunk_size=1000, request_timeout=200) but getting error elasticsearch.exceptions.RequestError: TransportError(400, u'action_request_validation_exception', u'Validation Failed: 1: index is missing;2: type is missing;3: index is missing;4: type is missing;5: index is missing;6: type is missing;7: index is missing;8)2020: this says