I’m not clear why an index on a hash map would be significantly smaller than an index on the same data in a cache?
Also, SQL has pretty good intersection/union operations; that’s not something you’d need to do yourself (unlike in Coherence). > On 18 Dec 2019, at 14:13, Rastogi, Arjit (CWM-NR) <arjit.rast...@rbccm.com> > wrote: > > Creating separate cache for HashMap may not be effective. > 1. The size of the index will be huge & it will impact fetch > performance. > 2. If we have to query on multiple entries in HashMap then we will have > to do individual queries & find intersection/ union result programmatically. > > Thanks & Regards, > Arjit Rastogi > > <>From: Stephen Darlington [mailto:stephen.darling...@gridgain.com] > Sent: Friday,December 13, 2019 4:44 PM > To: user@ignite.apache.org > Subject: Re: ValueExtractor support in Apache Ignite > > The “Ignite Way” of doing that would be to normalise the data, put the > IdentifierMap as a separate, probably co-located cache. > > Regards, > Stephen > > > On 13 Dec 2019, at 09:53, Rastogi, Arjit (CWM-NR) <arjit.rast...@rbccm.com > <mailto:arjit.rast...@rbccm.com>> wrote: > > Hi Ilya, > > We want to create index on the keys present in HashMap present in the POJO we > want to put in cache. > > Thanks & Regards, > Arjit Rastogi > From: Ilya Kasnacheev [mailto:ilya.kasnach...@gmail.com > <mailto:ilya.kasnach...@gmail.com>] > Sent: Thursday,December 12, 2019 4:13 PM > To: user@ignite.apache.org <mailto:user@ignite.apache.org> > Subject: Re: ValueExtractor support in Apache Ignite > > Hello! > > I don't think we have anything like that. > > Why do you need such baroque data structure? > > Regards, > -- > Ilya Kasnacheev > > > вт, 10 дек. 2019 г. в 13:39, Rastogi, Arjit (CWM-NR) <arjit.rast...@rbccm.com > <mailto:arjit.rast...@rbccm.com>>: > Hi All, > > What is the Oracle Coherence ValueExtractor > <https://docs.oracle.com/middleware/1213/coherence/java-reference/com/tangosol/util/ValueExtractor.html> > equivalent in Apache Ignite which can enable us to create dynamic indexes by > extracting elements in a Map/ List inside POJO. > > Use case: > Example: We want to create cache of following class- > class Employee { > long employeeId; > String employeeName; > Map<String, String> identifierMap; > } > > identifierMap example = [{“id1” : “123456”},{“id2” : “45678”}] > > We want to put objects of Employee in cache. We want to create index on all > entries in identifierMap like “id1”, “id2” etc. > We have achieved the same using ValueExtractor > <https://docs.oracle.com/middleware/1213/coherence/java-reference/com/tangosol/util/ValueExtractor.html> > in Oracle Coherence. > > Thanks & Regards, > Arjit Rastogi >