My table does't have indexes for punctuation. However, when I did a search on "1.23", it found the term but it took a while. Then I did a search on "Sandra O'connor", it didn't find the phrase even though it exists. Should'nt both of the phrases follow the same rule? Any help will be appreciated. Thanks.
Not really. The . and the digits cause 1.23 to to be considered non-language (see wordc and langc in the manual). Without the right index expression(s) it will just search for the 1 in the index then linearly for the rest. The ' is part of the language set so it tries to find it as is and doesn't because it's not indexed.
This is a useful expression to add using "addexp"
[\alnum\x80-\xff.]{1,70}>>[.&']=[\alnum\x80-\xff.]{1,70}
Also, before the language/non-language processing happens, the default index expression won't prefix-match `1.23', so the search would be entirely linear (no index could be used), which is why it would take a while.
Such a linear search would be disabled normally to save load; <apicp allinear on> enables it, but the better fix is to add the index expression as given so the search is faster.
For `"Sandra O'connor"', what is the exact character sequence that it occurs in the text, including a few surrounding characters? Also, what version of Texis is this (texis -version)?