Was there supposed to be a change in the behaviour of fuzzy matching between Elasticsearch 1.7 and 2.1?
A match query with fuzziness of "1" in Elasticsearch 2 will match transposed adjacent characters, but didn't in Elasticsearch 1.
E.g.
POST /foo/foo/1
{"text": "ab"}
GET /foo/foo/_search
{ "query": {"match": {"text": {"query": "ba", "fuzziness": 1}}}}
In Elasticsearch 1.7, the query did not match "ab", but it does match in 2.1