I think if you embed the sessionId into your value object and create an index, that should perform better. I am assuming if you embed the sesionId that you would then be doing queries more similar to sessionId = X or sessionId like sessionId%. Being able to use the index will prevent the query engine from having to do a full region scan.
For the like query, a % matcher preceding the value will also slow down the engine because it will have to scan a lot more of the region to find any possible matches. So %sessionId% will perform worse than sessionId%, especially with an index. Using PDX to serialize the value object would also help. On Sat, Sep 29, 2018 at 11:13 AM siby_sekar <siby_sek...@hotmail.com> wrote: > for us sessionId is not the actual key field. > > It will be something like some arbitrarystring+sessionid > > and we will be running queries with - like "%sessionId%" to find all > entries having a particular sessionid. > > In this case to improve performance can we apply any index ? > > is this approach better than running queries on value fields with PDX > Serialization ? > > > > -- > Sent from: > http://apache-geode-incubating-developers-forum.70738.x6.nabble.com/ >
