Reputation: 133
I have a collection which stores documents with the following structure:
{
"_id" : NumberLong(-6225637853094968071),
"Id" : "1585534",
"Type" : NumberInt(42),
"InDate" : ISODate("2017-10-05T18:30:00.000+0000"),
}
{
"_id" : NumberLong(-622563784353458071),
"Id" : "15832422",
"Type" : NumberInt(42),
"InDate" : ISODate("2017-10-06T00:00:00.000+0000"),
}
I want to delete all documents with a particular "Id" and "Type" where the "InDate" field has time 18:30:00.000+0000
. I tried using this query:
db.collection.remove({
"ChannelType": NumberInt(28),
"HotelId": "1585534",
"CheckInDate": /$18:30:00.000+0000/
})
but it doesnt work. How to do this?
Upvotes: 1
Views: 2641
Reputation: 151122
A Regular Expression will not work on something that is not a string. There actually is no simple way of selecting a BSON Date
by purely it's time value in a regular query expression.
In order to do this you would supply a $where
condition with the additional logic to match on just the time portion of the date value.
Based on the data actually supplied in the question:
db.getCollection('collection').remove({
"Id": "1585534",
"Type": NumberInt(42),
"$where": function() {
return this.InDate.getUTCHours() === 18
&& this.InDate.getUTCMinutes() === 30
}
})
Which is of course just the first document in the provided data, even though it's really just identified by "Id"
anyway since that value is unique between the two.
The .getUTCHours()
and .getUTCMinutes()
are JavaScript Date
object methods, which is how a BSON Date
is expressed when used within a $where
expression.
Trivially, MongoDB 3.6 ( upcoming as of writing ) allows a more efficient form than $where
in aggregation expressions:
db.getCollection('collection').aggregate([
{ "$match": {
"Id": "1585534",
"Type": NumberInt(42),
"$expr": {
"$and": [
{ "$eq": [{ "$hour": "$InDate" }, 18 ] },
{ "$eq": [{ "$minute": "$InDate" }, 30] }
]
}
}}
])
However such expressions cannot be used directly with methods such as .remove()
or .deleteMany()
, but $where
expressions can.
Upvotes: 2
Reputation: 2975
Iso date don't match a string, you need to use isodate in the query too
db.collection.remove({
"ChannelType": NumberInt(28),
"HotelId": "1585534",
"InDate": ISODate(2017-10-05T18:30:00.000+0000)
})
Upvotes: 1