AGGREGATE queries

The AGGREGATE query is used for performing aggregate() operations, via the aggregation framework.

The base syntax is: AGGREGATE <collection>. The remaining part will be used to build the aggregation pipeline.

Warning

This part is still very work-in-progress and things described below may not corespond to the actual state of the project.

Anyways, tests are being written in order to make things work this way.

The PROJECT command

Syntax: PROJECT <assignment-list>

The PROJECT command allows an “assignment list”, eg. a comma-separated list of SYMBOL = expression items, that will be used to build the map in the $project command.

Example:

AGGREGATE article
PROJECT title = 1,
        stats = {
            pv = '$pageViews',
            foo = '$other.foo',
            dpv = '$pageViews' + 10,
        }

results in:

db.article.aggregate({'$project': {
    'title': 1,
    'stats': {
        'pv': "$pageViews",
        'foo': "$other.foo",
        'dpv': {'$add': ["$pageViews", 10]}
    },
}})

The MATCH command

Syntax: MATCH <assignment-list>

Results in: {'$match': ...}

The LIMIT command

Syntax: LIMIT <num:int>

Results in: {'$skip': num}

The SKIP command

Syntax: SKIP <num:int>

Results in: {'$skip': num}

The UNWIND command

Syntax: UNWIND <name:symbol|string>

Results in: {'$unwind': name}

The GROUP command

Syntax: GROUP [ BY <expression> , ] <assignment-list>

Results in: {'$group': assignment_list}.

If the BY <expression> part is present, a key named _id will be added to the assignment list before using as argument to $group.

The SORT command

Syntax: SELECT <sort-spec>

Results in: {'$sort': sort_spec}.

Same as in the SELECT query: a list of <field> <direction> items will be converted in a list of (field, direction) tuples.

Supported directions are ASC and DESC.

The GEO_NEAR command

comments powered by Disqus