Hello!
If you like, you can use meaningful bits from the first two characters of
your key to form a partition id. You will need to inherit from
RendezvousAffinityFunction to do that.
In this case, you can just scan one partition (out of e.g. 1024).
However, this may pose a problem if data is
Another possibility is full-text search, which provides indexed search with
wildcards:
https://apacheignite.readme.io/v1.1/docs/cache-queries#text-queries
On Fri, Jan 31, 2020 at 2:29 PM Stephen Darlington <
stephen.darling...@gridgain.com> wrote:
> If you don’t use SQL, Ignite is basically a
If you don’t use SQL, Ignite is basically a key-value store. That is, if you
don’t know the key you have to look at every record to see if it matches.
You can specify a filter on the ScanQuery:
ScanQuery q = new ScanQuery<>((k,v) -> k.equals("Stephen"));
That wouldn’t be indexed, though. If
Sorry, I have not provided full information or not able to understand your
answer as i am newbie.
I am storing my keys and values as string.
For e.g. Key "ABC"+"id" value : serialized object in string form.
My objects/entities has no sql attributed like "QuerySqlField" and its not
possible
If you create an index on (A,B,C), SQL queries for all three variants you note
should work and use the index.
Having said that, “returning a huge number of rows” doesn’t seem like a good
usage pattern with Ignite. You might be better distributing your query around
the cluster rather than
I have key composed of "A"+"B"+"C"+Id.
I want to do search by "ABC*" or"AB*" or "A*".
Require very efficient query as i am fetching huge number of rows from cache
based on above query.
Also my values are serialized customize big user objects?
--
Sent from: