mongodb contains query empty result slow - mongodb

Question

I have around 10 millions document in MongoDB.
I'm trying to search for text inside the db db.outMessage.find({ "text" : /.*m.*/}) but it took too long (around 30 second) with no result, but if I search for existing text it took less than a second.
I tried to put index on text with same result.
db.outMessage.find({ "text" : /.*m.*/}).explain(true)
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "notification_center.outMessage",
"indexFilterSet" : false,
"parsedQuery" : {
"text" : {
"$regex" : ".*m.*"
}
},
"winningPlan" : {
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"filter" : {
"text" : {
"$regex" : ".*m.*"
}
},
"keyPattern" : {
"text" : 1
},
"indexName" : "text",
"isMultiKey" : false,
"multiKeyPaths" : {
"text" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"text" : [
"[\"\", {})",
"[/.*m.*/, /.*m.*/]"
]
}
}
},
"rejectedPlans" : [ ]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 0,
"executionTimeMillis" : 14354,
"totalKeysExamined" : 10263270,
"totalDocsExamined" : 0,
"executionStages" : {
"stage" : "FETCH",
"nReturned" : 0,
"executionTimeMillisEstimate" : 12957,
"works" : 10263271,
"advanced" : 0,
"needTime" : 10263270,
"needYield" : 0,
"saveState" : 80258,
"restoreState" : 80258,
"isEOF" : 1,
"invalidates" : 0,
"docsExamined" : 0,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"filter" : {
"text" : {
"$regex" : ".*m.*"
}
},
"nReturned" : 0,
"executionTimeMillisEstimate" : 12461,
"works" : 10263271,
"advanced" : 0,
"needTime" : 10263270,
"needYield" : 0,
"saveState" : 80258,
"restoreState" : 80258,
"isEOF" : 1,
"invalidates" : 0,
"keyPattern" : {
"text" : 1
},
"indexName" : "text",
"isMultiKey" : false,
"multiKeyPaths" : {
"text" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"text" : [
"[\"\", {})",
"[/.*m.*/, /.*m.*/]"
]
},
"keysExamined" : 10263270,
"seeks" : 1,
"dupsTested" : 0,
"dupsDropped" : 0,
"seenInvalidated" : 0
}
},
"allPlansExecution" : [ ]
},
"serverInfo" : {
"host" : "acsdptest.arabiacell.net",
"port" : 27017,
"version" : "3.4.7",
"gitVersion" : "cf38c1b8a0a8dca4a11737581beafef4fe120bcd"
},

Answer 1

The index will essentially be a list of all the values of the text field, in lexicographical order, i.e. sorted by the first letter.
Since the query executor has no way to predict which values might contain an 'm', it must examine all of the index entries.
In the case of this query, that means 10,263,270 index keys were examined, after being read from disk if the index was not already in the cache.
If this is actually a keyword search and not a single-letter match, instead of $regex, you might be able to make use of the $text query operator, which requires a text index