# Match data from two sets and mark entries **URL:** \ **Category:** Elasticsearch\ **Tags:** docker, painless\ **Created:** [August 28, 2021, 5:22pm UTC](https://discuss.elastic.co/t/match-data-from-two-sets-and-mark-entries/282731 "2021-08-28T17:22:38Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![mydeadvictoria](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mydeadvictoria/32/93843_2.png) [@mydeadvictoria](https://discuss.elastic.co/u/mydeadvictoria)\ **Post date:** [August 28, 2021, 5:22pm UTC](https://discuss.elastic.co/t/match-data-from-two-sets-and-mark-entries/282731/1 "2021-08-28T17:22:38Z")
I have two relatively large sets of people data (first\_name, last\_name, birth\_date etc). And I need to 'match' them, because some people are present in both sets and I need to find and mark them (by adding their ID from the other set). I've come up with this solution: 1. Load set A into Elasticsearch index (6M entries) 2. During load of set B, for each entry, make an update-by-query request which looks for people with the same first\_name(Text), last\_name(Text) & birth\_date(Date) and adds `B_id` field through a simple script (painless) So for example, `A` index roughly might look like this: | \_id | first\_name | last\_name | birth\_date | b\_id | | --- | --- | --- | --- | --- | | yJzdWIiOiJ | demo1 | qwerty | 2001.11.04 | 4d5323bd91c2 | | MKPj0ILgq1 | demo2 | demo2 | 1995.11.11 | null | | oueUg3sBO | demo512 | demo512 | 2000.05.16 | null | Here, an entry with id `yJzdWIiOiJ` got matched with an entry with id `4d5323bd91c2` from index `B`. An example request for such flow: ```json { "script":{ "id":"id-field-adding-script", "params":{ "value":9150456064 } }, "size":1000, "query":{ "bool":{ "must":[ { "match":{ "first_name":{ "query":"Ryan", "operator":"AND", "prefix_length":0, "max_expansions":50, "fuzzy_transpositions":true, "lenient":false, "zero_terms_query":"NONE", "auto_generate_synonyms_phrase_query":true, "boost":1.0 } } }, { "match":{ "last_name":{ "query":"Jewel", "operator":"AND", "prefix_length":0, "max_expansions":50, "fuzzy_transpositions":true, "lenient":false, "zero_terms_query":"NONE", "auto_generate_synonyms_phrase_query":true, "boost":1.0 } } }, { "match":{ "birth_date":{ "query":"1992-09-11", "operator":"OR", "prefix_length":0, "max_expansions":50, "fuzzy_transpositions":true, "lenient":false, "zero_terms_query":"NONE", "auto_generate_synonyms_phrase_query":true, "boost":1.0 } } } ], "adjust_pure_negative":true, "boost":1.0 } } } ``` This request is sent for every entry (via Java Spring backend). it is not as slow as I expected it to be, but I still want to improve it. In my tests, set `A` has 6M entries and set `B` has ~250K. It takes up to 2 hours to both match set `B` with set `A` and import set `B` afterwards (actually it takes 10k entries, matches them and then imports into ES, repeats). **The question is** : How do I improve this, make it faster? Or maybe you can suggest another approach. ---
**Author:** ![spinscale](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/spinscale/32/25011_2.png) [@spinscale](https://discuss.elastic.co/u/spinscale)\ **Post date:** [August 30, 2021, 11:16am UTC](https://discuss.elastic.co/t/match-data-from-two-sets-and-mark-entries/282731/2 "2021-08-30T11:16:21Z")
There is no built-in function to solve this very concrete use-case. You could possibly do this with two single searches.. albeit longer ones 🙂 How about running two scroll/PIT searches, one against index a, one against index b. Sort by last name, first name, birth\_date and then iterate through the index with less documents and see compare to the index with more indices, when you find a matching document. The idea would be something like leap-frogging/zig zag joining the data and thus find the two matches. Whenever you find a match, run an update query (and potentially bulk those to save some more resources). Does that make sense? ---
**Author:** ![mydeadvictoria](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/mydeadvictoria/32/93843_2.png) [@mydeadvictoria](https://discuss.elastic.co/u/mydeadvictoria)\ **Post date:** [August 30, 2021, 4:03pm UTC](https://discuss.elastic.co/t/match-data-from-two-sets-and-mark-entries/282731/3 "2021-08-30T16:03:25Z")
I'm relatively new to ES, not sure I get the idea. Could you please show some example? ---
**Author:** ![vincenbr](https://avatars.discourse-cdn.com/v4/letter/v/8edcca/32.png) [@vincenbr](https://discuss.elastic.co/u/vincenbr)\ **Post date:** [August 30, 2021, 5:07pm UTC](https://discuss.elastic.co/t/match-data-from-two-sets-and-mark-entries/282731/4 "2021-08-30T17:07:06Z")
When I hear "match data" in elasticsearch, the [percolator](https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-percolate-query.html) automatically comes to my mind... I can't dig into it further now, but maybe index set B in a percolation index (with term queries), and percolate documents from set A into it, in order to retrieve matching ids ? ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [September 27, 2021, 5:07pm UTC](https://discuss.elastic.co/t/match-data-from-two-sets-and-mark-entries/282731/5 "2021-09-27T17:07:56Z")
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.