How to say… match when field is a number… in mongodb?

问题

So I have a field called 'city' in my results...the results are corrupted, some times it's an actual name, sometimes it's a number. The following code displays all the records...

db.zips.aggregate([{$project : {city:{$substr:["$city",0,1]}}},{$sort : {city : 1}} ])

I need to modify this line to display only the records with a city who has a name that is a number (2,3,4, etc)....I think I can use '$match', but how?

db.zips.aggregate([{$project : {city:{$substr:["$city",0,1]}}},{$sort : {city : 1}}, {$match:{???what_to_say_here???} ])

How to say 'match when city is a number'?

the out put I get looks like this...

    {
        "city" : "A",
        "_id" : "04465"
    },
    {
        "city" : "1",
        "_id" : "02821"
    },
    {
        "city" : "0",
        "_id" : "04689"
    }

I'm trying to display only the records with a numeric string...this is related to a larger "homework" problem but I can't even get to the actual homework question until I get past this point.

回答1:

Use the $type operator in your $match:

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]}}},
    {$sort : {city : 1}}, 
    {$match: {city: {$type: 16}}}      // city is a 32-bit integer
]);

There isn't a single type value for number so you need to know which type of number you have:

32-bit integer   16
64-bit integer   18
Double           1

Or use an $or operator to match all types of numbers:

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]}}},
    {$sort : {city : 1}}, 
    {$match: {$or: [{city: {$type: 1}}, {city: {$type: 16}}, {city: {$type: 18}}]}}
]);

Or even use $not to match all docs where city is not a string:

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]}}},
    {$sort : {city : 1}}, 
    {$match: {city: {$not: {$type: 2}}}}      // city is not a string
]);

UPDATED

To match all docs where city is a numeric string you can use a regular expression:

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]}}},
    {$sort : {city : 1}}, 
    {$match: {city: /^\d.*$/}}      // city is all digits
]);

回答2:

Why not to use $regex?

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]}}},
    {$sort : {city : 1}}, 
    {$match: {city:{$regex:'[0-9]'}}}
])

回答3:

Simply use:

db.zips.aggregate([
{$match: {
    'city': { $regex: '^[0-9].*'}
}}])

This work fine for me!

回答4:

An easier way to match all docs where city is a numeric string is using the property '0' <= city <= '9' i.e. your aggregation query becomes:

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]}}},
    {$sort : {city : 1}}, 
    {$match: {city: {$gte: '0', $lte: '9'}}}
]);

When using JohnnyHK's suggestion of a regex you can save yourself the $project stage altogether.

回答5:

You can also use an in command

db.zips.aggregate([     
    {
        $project: {
            first_char: { 
                $substr : ["$city",0,1]
            }, 
            "_id":"$_id"
        } 
    }, 
    {
        $match: {
            first_char: { 
                $in: ["0","1","2","3","4","5","6","7","8","9"] 
            }
        }
    }
])

回答6:

you can also try like the following, using regular expression without using $regex

db.zips.aggregate([
{$project : { city : { $substr : ["$city",0,1] } }}, 
{$sort : {city : 1}}, 
{$match: { city:/[0-9]/}}
])

回答7:

db.zips.aggregate([
    {$project : {city:{$substr:["$city",0,1]},pop:1}},
    {$group:{_id:"$city",sumPop:{$sum:"$pop"}}},
    {$sort : {_id : 1}}, 
    {$match:{_id : {$regex:/\d/}} },
    {$group:{_id:"id",sumPop:{$sum:"$sumPop"}}},
]);

回答8:

With the mongodb 4.4 (upcoming), You can use $isNumber pipeline operator to check whether the expression is integer or other BSON type.

db.zips.find({ "$expr": { "$anyElementTrue": { "$isNumber": "$city" }}})

$isNumber will return true or false corrsoponding to the expression and $anyElementTrue will return the $matched documents.

来源：https://stackoverflow.com/questions/13409386/how-to-say-match-when-field-is-a-number-in-mongodb

标签

python

mongodb