AFAIK, for MongoDB plugin, we have very limited tests. Anyway, I tried the following query on your sample json file:
0: jdbc:drill:schema=dfs.tpchPar100> select t1.c1.text, count(t1.c1.text) from (select flatten(tbl.twitter_entities.hashtags) as c1 from dfs.`/jsondata/junk.json` as tbl) t1 group by t1.c1.text; +---------+---------+ | EXPR$0 | EXPR$1 | +---------+---------+ | skiing | 1 | | snow | 1 | +---------+---------+ 2 rows selected (0.542 seconds) is this the group/count you want to retrieve? HTH, Dechang On Sun, Dec 11, 2016 at 12:10 AM, yousuf <[email protected]> wrote: > Hi, > > How can I retrieve hashtags[n].text values as flatten from the following > json document (storage mongo). My final goal is to group text with count. > > Apache drill version : 1.8.0 > MongoDB version: 3.2 > > I've tried several examples but my final result is either empty array eg > [] or blank. > > { > "twitter_entities": { > "hashtags": [{ > "text": "snow", > "indices": [ > 66, > 71 > ] > }, { > "text": "skiing", > "indices": [ > 73, > 80 > ] > }], > "trends": [], > "urls": [{ > "url": "http://xxxxxxx";, > "expanded_url": "http://xxxxx";, > "display_url": xxxxx", > "indices": [ > 71, > 93 > ] > }] > } > } > > Any help is appreciated > Regards > Yousuf > > >
