preserve_original
and preserve_original_recorder
are token filters that allow to keep and index original terms. This is very similar to the keyword_repeat
and unique
filters but will work also on filters that do not support the keyword attribute.
index : analysis : analyzer : preserve_case : type : custom tokenizer : whitespace filter : [preserve_original_recorder, lowercase, preserve_original]
Will produce the following terms hello
, Hello
, the
, world
, World
for the input text Hello the World
.