mongo-aggregation: apply regex grouping, string processing on $project

Question

I'd like to apply some simple String manipulation when doing $project, is it possible to apply something like the following function on $project? :

var themeIdFromZipUrl = function(zipUrl){
    return zipUrl.match(/.*\/(T\d+)\/.*/)[1]
};

I'm using the following query:

db.clientRequest.aggregate(
{
$match: {
  "l": {$regex: ".*zip"},
  "t": { "$gte": new Date('1/SEP/2013'),
                    "$lte": new Date('7/OCT/2013')
                    }
  }
},
{
  $project: {"theme_url" : "$l", "_id": 0, "time": "$t"}
},
{
  $group: {   _id: {
                      theme_url: "$theme_url",
                      day: {
                              "day": {$dayOfMonth : "$time"},
                              "month": {$month: "$time"},
                              "year": {$year: "$time"}
                            },
              },
              count: {$sum:1}
  }
}

)

This returns following:

        {
        "_id" : {
            "theme_url" : "content/theme/T70/zip",
            "day" : {
                "day" : 13,
                "month" : 9,
                "year" : 2013
            }
        },
        "count" : 2
    }

Can I apply the function above on the theme_url field and turn it to theme_id? I took a little look on Map-Reduce, but I'm not sure whether it's a bit too complicated for such an easy case.

Thanks,

Amit.

there isn't currently support for doing this directly but there is an ugly way to do it with aggregation framework (involves doing character comparisons to figure out where to cut the string). — Asya Kamsky
– Asya Kamsky, Commented Jan 11, 2014 at 9:39

Avish · Accepted Answer · 2014-01-28 07:18:06Z

2

There's no way to do this using the Aggregation Framework currently.

You could do it with MapReduce but that would probably slow down the entire thing (if the amount of data is large).

If this is the last step of the aggregation you can also do it on the clientside after the aggregation completes. e.g. in the Mongo shell:

var aggregationResults = col.aggregate([ /* aggregation pipeline here */]);
aggregationResults.results.forEach(function(x) { 
  x._id.theme_id = themeIdFromUrl(x._id.themeUrl);
});

If you're using a driver for another language you'll have to do this in whatever language you're using, of course.

Generally speaking, if your data contains a theme_url and the theme_id is encoded in the URL, it might make sense to store it in its own field. Mongo is not a very good tool for text manipulation.

answered Jan 28, 2014 at 7:18

Avish

4,64622 silver badges28 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

mongo-aggregation: apply regex grouping, string processing on $project

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related