ElasticSearch[5.5] Aggs Range filter on string datatype

Elastic Stack Elasticsearch
1

RAM field in Mobile specifications has lots of data which i need to apply some kind of Range filter on strings . I have formatted bucket into key value pair RAM and Doc count
"RAM": {
"1": 7,
"32": 1,
"32 MB": 490,
"1 GB": 426,
"2 GB": 361,
"3 GB": 269,
"4 GB": 143,
"512 MB": 136,
"256 MB": 87,
"10 MB": 61,
"64 MB": 47,
"1 MB": 45,
"128 MB": 41,
"56 MB": 27,
"1.5 GB": 23,
"16 MB": 12,
"35 MB": 12,
"4 MB": 11,
"0.032 MB": 10,
"100 MB": 7,
"3 MB": 7,
"2 MB": 6,
"768 MB": 6,
"40 MB": 4,
"30 MB": 3,
"6 GB": 3,
"20 MB": 2,
"46 MB": 2,
"500 MB": 2,
"60 MB": 2,
"120 MB": 1,
"131 MB": 1,
"138 MB": 1,
"153 MB": 1,
"200 MB": 1,
"24 MB": 1,
"250 MB": 1,
"260 MB": 1,
"50 MB": 1,
"55 MB": 1,
"65 MB": 1,
"8 GB": 1
}

How to convert the above one to something like below

 128 MB and Below
     1 GB
     2 GB
     3 GB
     4 GB
     6 GB and Above

One possible solution would be index the value of RAM as numeric datatype and apply range filter on the field. But how to differentiate 1 Mb and 1 GB .

2

If you convert all that in the lowest unit you can have, let say kb like 1mb = 1024 and 1gb = 1024*1024 for example, then it would become obvious.

My 2 cents.

That said it could be nice that we support (in core or as a plugin) a « size » type with size_histogram agg and size_range agg.
I did not check if that has been already asked though.

3

That could be a good solution for RAM field and what about incase of field like battery capacity.

It would be good if it has inbuilt support like size aggregate as you mentioned.

4

That’s the same for whatever unit. If you have something in W, kW, mW, just Store it as Watts and you re done

5

Thanks this work perfectly. I have stored data without and with conversation.

6

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.