This is the expected/intended behavior for an indexed field. This behavior (as well as fuzzy and root variation) is why we don't recommend indexing ID-like fields.

Indexed searches add fuzzy and root word variation, as you noted, but what you are seeing is that they also perform a search within the text of a field.

Think of when you search on document text: if you want to return a document with the word "Laserfiche" in it, you perform a search for the word Laserfiche with no wildcards, and it returns all documents where the word appears ANYWHERE within the document text.

This acts the same way. Without fuzzy search and root word variation, the results will be the same as a non-indexed search with leading and trailing wildcards --- in your case, *1009*.

One final piece to answer your puzzle: when indexing, all punctuation is treated as spaces. That is,  1009-001  is the same as 1009 001 for an indexed search.

As a workaround, you can change the syntax generated by your search to force a non-indexed search. You should see a ~= instead of just a = in your search syntax, such as {[yourTemplate]:[yourField]~="1009"} . If you remove the ~, so that you have {[yourTemplate]:[yourField]="1009"}, a non-indexed search will be performed.