The letter tokenizer discards all non-letter characters from the input string and then generates a token at strings of contiguous letters.

Factory class: solr.LetterTokenizerFactory

Arguments: None

Example:

<fieldType name="text_en" class="solr.TextField" positionIncrementGap="100">
 <analyzer>
 <tokenizer class="solr.LetterTokenizerFactory"/>
 </analyzer>
</fieldType>

Input: I haven't received mail by Nov12Sunday

Output: I, haven, t, received, mail, by, Nov, Sunday

All non-letter characters (' and 12) are discarded first, and then tokens are generated by considering strings of contiguous letters.