Notes

# from datetime import datetime
from elasticsearch import Elasticsearch
from elasticsearch_dsl import UpdateByQuery
import elasticsearch_dsl
import elasticsearch


es = Elasticsearch([{'host': 'localhost', 'port':'9200'}])
es_index = 'customer_all_index'
grp_name_query = {
"aggs": {
"GroupByType": {
"terms": {
"field": "FirstName.first_key"
},
"aggs": {
"Group": {
"top_hits": {
"size": 12, # will return at most records
"_source": {
"include": ['Email', "CustomerNo", "FirstName", "LastName", "Phoneno", "MasterID"]
},
"sort": [{ # will sort by lowest value first
"CustomerNo.keyword": {
"order": "asc"
}
}]
}
}
}
}
}
}


def update_child(doc_id, min_master_id):

q = {
"doc": {
"MasterID": min_master_id
}
}
es.update(index=customer_all_index, body=q, id=doc_id)
print('Updated')


def update_parent(doc_id, min_master_id):

q = {
"doc": {
"MasterID": min_master_id
}
}
es.update(index=customermerged_15, body=q, id=doc_id)
print('Updated')

max_merged_assigned_query={ "size":1,
"query": {
"match_all": {}
},
"sort": [
{
"MasterID": {
"order": "desc"
}
}

]

}

results1=es.search(index=es_index,body=max_merged_assigned_query)

max_merged_assigned= results1['hits']['hits'][0]['_source']['MasterID']
print(max_merged_assigned)


results2 = es.search(index='customer_all_index', body=grp_name_query, size=15)

for bucket2 in results2['aggregations']['GroupByType']['buckets']:
temp_mergeID=[]
for hit2 in bucket2['Group']['hits']['hits']:
doc_id= hit2['_id']
temp_mergeID.append( int(hit2['_source']['MasterID']))
print(hit2['_source']['MasterID'])
print(temp_mergeID)
min_master_id=int(hit2['_source']['MasterID'])
if(min_master_id>hit2['_source']['MasterID']):
min_master_id=hit2['_source']['MasterID']
#update_child(doc_id,min_master_id)
#update_parent(doc_id,min_master_id)
print("min "+str(min_master_id))
# update_parent_child(temp_mergeID)
print("updated")