i have collection stats.daily stores multiple dimensional data, depending type requires other specific keys.
, it's required these keys (depending type) unique.
example, in case url key required because type "url":
{ "site": 1, "type": "url", "url": "http://google.com/" "totals": { "variable1": 12, // incrementing values "variable2": 32 } } another example:
{ "site": 1, "type": "domain", "domain": "google.com", "totals": {...} } so create indexes:
db.coll.createindex({site: 1, type: 1, url: 1}, {unique: true, sparse: true}); db.coll.createindex({site: 1, type: 1, domain: 1}, {unique: true, sparse: true}); it doesn't works, returns exception: e11000 duplicate key error index. makes sense unique index, should work sparse part.
what's best solution accomplish need?
edit:
in cases, depending type, there may have more 1 related key:
{ "site": 1, "type": "search", "engine": "google", "term": "programming", "totals": {...} } and unique index include these 2 new keys.
Comments
Post a Comment