elasticsearch/docs/reference/migration/migrate_3_0.asciidoc

[[breaking-changes-3.0]]
== Breaking changes in 3.0

This section discusses the changes that you need to be aware of when migrating
your application to Elasticsearch 3.0.

=== Search changes

==== `search_type=count` removed

The `count` search type was deprecated since version 2.0.0 and is now removed.
In order to get the same benefits, you just need to set the value of the `size`
parameter to `0`.

For instance, the following request:
[source,sh]
---------------
GET /my_index/_search?search_type=count
{
  "aggs": {
    "my_terms": {
       "terms": {
         "field": "foo"
       }
     }
  }
}
---------------

can be replaced with:
[source,sh]
---------------
GET /my_index/_search
{
  "size": 0,
  "aggs": {
    "my_terms": {
       "terms": {
         "field": "foo"
       }
     }
  }
}
---------------

==== `search_type=scan` removed

The `scan` search type was deprecated since version 2.1.0 and is now removed.
All benefits from this search type can now be achieved by doing a scroll
request that sorts documents in `_doc` order, for instance:

[source,sh]
---------------
GET /my_index/_search?scroll=2m
{
  "sort": [
    "_doc"
  ]
}
---------------

Scroll requests sorted by `_doc` have been optimized to more efficiently resume
from where the previous request stopped, so this will have the same performance
characteristics as the former `scan` search type.

=== Parent/Child changes

The `children` aggregation, parent child inner hits and `has_child` and `has_parent` queries will not work on indices
with `_parent` field mapping created before version `2.0.0`. The data of these indices need to be re-indexed into a new index.

The format of the join between parent and child documents have changed with the `2.0.0` release. The old
format can't read from version `3.0.0` and onwards. The new format allows for a much more efficient and
scalable join between parent and child documents and the join data structures are stored on on disk
data structures as opposed as before the join data structures were stored in the jvm heap space.

==== `score_type` has been removed

The `score_type` option has been removed from the `has_child` and `has_parent` queries in favour of the `score_mode` option
which does the exact same thing.

==== `sum` score mode removed

The `sum` score mode has been removed in favour of the `total` mode which doesn the same and is already available in
previous versions.

==== `max_children` option

When `max_children` was set to `0` on the `has_child` query then there was no upper limit on how many children documents
are allowed to match. This has changed and `0` now really means to zero child documents are allowed. If no upper limit
is needed then the `max_children` option shouldn't be defined at all on the `has_child` query.