Hi All,
Is there any possibility to create a unique index number while indexing because i want to search the result on behalf of last event id i read.then i will read the event greater then from that last event id.
The problem i faced is device is generating wrong date time and that time is future date time in that case our earliest and latest logic is failing. because i was storing the _time value as a last trigger date and that was future date.so i want to create a unique auto incremented row number id while indexing.
Please help!!
You could use _indextime
instead of _time
. If that's not enough you could also use (or combine) the _cd
field for uniqueness. Both are a kind of incrementing counters.
_cd
is a unique number within an index (it consists of a bucket number and an offset within that bucket). It always exists. It does not increment just by 1 though, the only guarantee you have is that it's incrementing.
Not sure what you mean by that it's taking a long time to search and why.
Thanks.. can we get _cd in every type of device logs?
Can we get the two ranges of _cd logs e.g i got 1234 in _cd now i want the logs between 1234 and 1240 o/p = 6 events?
_indextime is returning epoch time and its taking long time i search.
You could use _indextime
instead of _time
. If that's not enough you could also use (or combine) the _cd
field for uniqueness. Both are a kind of incrementing counters.