With the CSV test dataset, the following queries return columns in different order
from message_types,languages_lookup_non_unique_key
| sort country.keyword NULLS FIRST
| eval language_code = 1, `language_name` = false, message = true, foo_3 = 1, foo_2 = null
| eval foo_3 = "1", `foo_3` = -1, foo_1 = 1, `language_code` = null, `foo_2` = "1"
| mv_expand foo_1
country | country.keyword | type | language_name | message | foo_3 | foo_1 | language_code | foo_2
------------------------+------------------------+---------------+---------------+---------------+---------------+---------------+---------------+---------------
null |null |null |false |true |-1 |1 |null |1
...
same query, but with a LIMIT
from message_types,languages_lookup_non_unique_key
| sort country.keyword NULLS FIRST
| eval language_code = 1, `language_name` = false, message = true, foo_3 = 1, foo_2 = null
| eval foo_3 = "1", `foo_3` = -1, foo_1 = 1, `language_code` = null, `foo_2` = "1"
| mv_expand foo_1
| limit 7117
country | country.keyword | type | language_name | message | foo_3 | language_code | foo_2 | foo_1
------------------------+------------------------+---------------+---------------+---------------+---------------+---------------+---------------+---------------
null |null |null |false |true |-1 |null |1 |1
...
See the difference in the last three columns:
foo_1 | language_code | foo_2
vs
language_code | foo_2 | foo_1
With the CSV test dataset, the following queries return columns in different order
same query, but with a
LIMITSee the difference in the last three columns:
vs