<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-cite-prefix">On 4/14/25 23:03, Fred wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CAF2xExu8JQkMdenSbbY+nwt4LzqrQb88bO7EU6OZqM1AzMC+kw@mail.gmail.com">do
I need to open a Sword bug?</blockquote>
<br>
<font face="FreeSerif">If one is needed, I'm happy to do so. But
further experimentation using diatheke was educational. Now I have
more understanding but less explanation. And the difference is
clearly happening in Sword, not Xiphos.<br>
<br>
Xiphos' sidebar search, in the lucene case, reworks the search
request to convert each space-delimited TERM to +TERM so that
lucene search always asks for a logical "and" of the TERMs
offered. "and" is the normal use case for simple sidebar search.
Xiphos does this even when there's just one word, which should
(hm, "should") make it redundant but harmless. I've found that the
difference is found between<br>
</font><font face="monospace">diatheke -b 2TGreek -s lucene -k </font><font
face="monospace">παντοκρατωρ</font><font face="FreeSerif"><br>
and<br>
</font><font face="monospace">diatheke -b 2TGreek -s lucene -k
+παντοκρατωρ</font><font face="FreeSerif"><br>
The latter returns nothing, the former returns the expected set.
The same holds true in e.g. TischMorph and TR.<br>
<br>
I looked at the difference between Xiphos' sidebar and advanced
search to confirm this: Advanced search does no '+' favors, it
just takes the user's input literally. So when I do advanced
search for παντοκρατωρ, I get a reasonable set, but when I do so
using +παντοκρατωρ, I get nothing.<br>
<br>
Troy, can you offer an explanation for why (some) single term
searches have no results when prefixed with '+'? This doesn't
happen in all cases, e.g. I can search for "jesus" alone in any
English bible, with or without leading '+', and get the same
results.<br>
<br>
</font><font face="monospace">diatheke -b NET -s lucene -k +jesus |
tr ';' '\n' | wc -l<br>
1157<br>
diatheke -b NET -s lucene -k jesus | tr ';' '\n' | wc -l<br>
1157<br>
diatheke -b NASB -s lucene -k +jesus | tr ';' '\n' | wc -l<br>
948<br>
diatheke -b NASB -s lucene -k jesus | tr ';' '\n' | wc -l<br>
948<br>
diatheke -b KJV -s lucene -k +jesus | tr ';' '\n' | wc -l<br>
943<br>
diatheke -b KJV -s lucene -k jesus | tr ';' '\n' | wc -l<br>
943<br>
</font><font face="FreeSerif"><br>
All that said, I still have no explanation at all for how some
results in Matt, Mark, 1Pet etc are being obtained for regex or
phrase searches.<br>
</font>
</body>
</html>