thank you for your suggestions, Michael. I am going to start testing with using a nested object for my retailer data and then also disabling the source storage of the original JSON. Will post my results.
On Saturday, February 15, 2014 9:35:10 AM UTC-5, Michael Sokolov wrote: > > The problem with the parent/child thing is that each child can only have a > single parent. So if you index what -- in your model -- is the same > product as a child of multiple vendors, and then you want to search for > multiple vendors at once, you are going to get multiple copies of the child > product, one for each vendor that has it. I think probably if you were > always searching for a single vendor at a time, this could work though. > > Indexing all the vendor ids in each product document probably won't add > any indexing overhead relative to the parent/child model you proposed -- in > both cases you have to reindex the products when a vendor picks them up or > drops them, although in the parent/child case ES does this for you > internally. > > Re: the huge documents; yes that could be a problem. Although having lots > of terms in the index is not really an issue, the fact that ES stores your > entire document and gives it back to you by default will chew up storage > and bandwidth, as you say. To avoid this, you could disable the _source > field: > http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/mapping-source-field.html > > -Mike > > > PS I got your first reply - it just didn't go to the list > > On Friday, February 14, 2014 11:34:11 PM UTC-5, Ben Hirsch wrote: >> >> argh I just wrote a reply but google ate it apparently. >> >> So, are you suggesting that scrapping parent/child and simply storing all >> of the retailer data in the product document is a safer bet. I imagine we >> could rate limit our product indexing. However this now gives me two >> concerns: 1 - the size of the product document. We would have product >> documents with 20,000+ UPC entries as nested objects! and 2 - would our >> search results also return those 20,000 nested objects? >> > -- You received this message because you are subscribed to the Google Groups "elasticsearch" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/1e0b59fb-0514-40a4-8e8d-20f807aabcd0%40googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
