如何在Python中使用发布请求在Elasticsearch中建立索引？

Question

I am facing an issue:

RequestError(400, 'illegal_argument_exception', 'mapper [columns.analysis.abstract_stats.description.std] of different type, current_type [text], merged_type [float]')

which led me to go for a solution described here .

My current code which is generating the aforementioned error is:

from test_mapping import a

es = Elasticsearch([{'host': 'A.B.C.D', 'port': 9200}])

try:
    es.index(index='datatables', doc_type='datatable_v1', id="pallet_d3dd6729b810bebd955708e85afc1f65c3f2685c", body=a)
except Exception as e:
    print (e)

The index existed before but I have deleted it and then running the above code is still generating the above error. The variable a is here

Answer 1

The reason for the above error is when you send data to elastic then it created dynamic field for the keys it find missing in the mapping and try to identify its type. Base on the data you are sending in body the value at columns.analysis.abstract_stats.description.std is mapped to float type but one of the record at key columns.analysis.abstract_stats.description.std has value 'NaN' which can't be mapped to a float field and hence the error. You need to make sure that type of fields doesn't change from one record to another.

Answer 2

Use this to load a:

import simplejson
es.index(index='datatables', doc_type = 'datatable_v1', id = "pallet_d3dd6729b810bebd955708e85afc1f65c3f2685c", body = simplejson.dumps(a, ignore_nan = True))

This should solve your problem. Now your application will read this value as None (which is possibly the source of this corruption) and you can easily implement your functionality