$pull¶

On this page

$pull¶

The $pull operator removes from an existing array all instances of a value or values that match a specified condition.

The $pull operator has the form:

{ $pull: { <field1>: <value|condition>, <field2>: <value|condition>, ... } }

To specify a <field> in an embedded document or in an array, use dot notation.

Behavior¶

If you specify a <condition> and the array elements are embedded documents, $pull operator applies the <condition> as if each array element were a document in a collection. See Remove Items from an Array of Documents for an example.

If the specified <value> to remove is an array, $pull removes only the elements in the array that match the specified <value> exactly, including order.

If the specified <value> to remove is a document, $pull removes only the elements in the array that have the exact same fields and values. The ordering of the fields can differ.

Examples¶

Remove All Items That Equal a Specified Value¶

Given the following document in the stores collection:

{
   _id: 1,
   fruits: [ "apples", "pears", "oranges", "grapes", "bananas" ],
   vegetables: [ "carrots", "celery", "squash", "carrots" ]
}
{
   _id: 2,
   fruits: [ "plums", "kiwis", "oranges", "bananas", "apples" ],
   vegetables: [ "broccoli", "zucchini", "carrots", "onions" ]
}

The following operation updates all documents in the collection to remove "apples" and "oranges" from the array fruits and remove "carrots" from the array vegetables:

db.stores.update(
    { },
    { $pull: { fruits: { $in: [ "apples", "oranges" ] }, vegetables: "carrots" } },
    { multi: true }
)

After the operation, the fruits array no longer contains any "apples" or "oranges" values, and the vegetables array no longer contains any "carrots" values:

{
  "_id" : 1,
  "fruits" : [ "pears", "grapes", "bananas" ],
  "vegetables" : [ "celery", "squash" ]
}
{
  "_id" : 2,
  "fruits" : [ "plums", "kiwis", "bananas" ],
  "vegetables" : [ "broccoli", "zucchini", "onions" ]
}

Remove All Items That Match a Specified `$pull` Condition¶

Given the following document in the profiles collection:

{ _id: 1, votes: [ 3, 5, 6, 7, 7, 8 ] }

The following operation will remove all items from the votes array that are greater than or equal to ($gte) 6:

db.profiles.update( { _id: 1 }, { $pull: { votes: { $gte: 6 } } } )

After the update operation, the document only has values less than 6:

{ _id: 1, votes: [  3,  5 ] }

Remove Items from an Array of Documents¶

A survey collection contains the following documents:

{
   _id: 1,
   results: [
      { item: "A", score: 5 },
      { item: "B", score: 8, comment: "Strongly agree" }
   ]
}
{
   _id: 2,
   results: [
      { item: "C", score: 8, comment: "Strongly agree" },
      { item: "B", score: 4 }
   ]
}

The following operation will remove from the results array all elements that contain both a score field equal to 8 and an item field equal to "B":

db.survey.update(
  { },
  { $pull: { results: { score: 8 , item: "B" } } },
  { multi: true }
)

The $pull expression applies the condition to each element of the results array as though it were a top-level document.

After the operation, the results array contains no documents that contain both a score field equal to 8 and an item field equal to "B".

{
   "_id" : 1,
   "results" : [ { "item" : "A", "score" : 5 } ]
}
{
  "_id" : 2,
  "results" : [
      { "item" : "C", "score" : 8, "comment" : "Strongly agree" },
      { "item" : "B", "score" : 4 }
   ]
}

Because $pull operator applies its query to each element as though it were a top-level object, the expression did not require the use of $elemMatch to specify the condition of a score field equal to 8 and item field equal to "B". In fact, the following operation will not pull any element from the original collection.

db.survey.update(
  { },
  { $pull: { results: { $elemMatch: { score: 8 , item: "B" } } } },
  { multi: true }
)

However, if the survey collection contained the following documents, where the results array contains embedded documents that also contain arrays:

{
   _id: 1,
   results: [
      { item: "A", score: 5, answers: [ { q: 1, a: 4 }, { q: 2, a: 6 } ] },
      { item: "B", score: 8, answers: [ { q: 1, a: 8 }, { q: 2, a: 9 } ] }
   ]
}
{
   _id: 2,
   results: [
      { item: "C", score: 8, answers: [ { q: 1, a: 8 }, { q: 2, a: 7 } ] },
      { item: "B", score: 4, answers: [ { q: 1, a: 0 }, { q: 2, a: 8 } ] }
   ]
}

Then you can specify multiple conditions on the elements of the answers array with $elemMatch:

db.survey.update(
  { },
  { $pull: { results: { answers: { $elemMatch: { q: 2, a: { $gte: 8 } } } } } },
  { multi: true }
)

The operation removed from the results array those embedded documents with an answers field that contained at least one element with q equal to 2 and a greater than or equal to 8:

{
   "_id" : 1,
   "results" : [
      { "item" : "A", "score" : 5, "answers" : [ { "q" : 1, "a" : 4 }, { "q" : 2, "a" : 6 } ] }
   ]
}
{
   "_id" : 2,
   "results" : [
      { "item" : "C", "score" : 8, "answers" : [ { "q" : 1, "a" : 8 }, { "q" : 2, "a" : 7 } ] }
   ]
}