Bug: Results of Wildcard Search in OT-Quotations in HCSB

For the genesis of this problem see http://community.logos.com/forums/p/26181/193398.aspx#193398
(I was using 4.2 B7).

The difference in results 4322 vs 4588 in the same number of verses prompted me check the search results:-

The red lined words are part of the quotation in the bible. I get exactly the same results with 4.1 SR-4 on my desktop and laptop. In an attempt to fix the anomaly:-

I rebuilt both Bible and Library Indexes in 4.2
forced the the server to re-download HCSB.logos4 (it was exactly the same)

I still get 4322 results.

What is the right result?

Why are parts of the search results not highlighted?

Find more posts tagged with

Comments

Melissa Snyder

I get 4,322 as well. Odd that pretty much the same words are not highlighted in Dominick's and Mark's 4,588 results. I'll submit a bug report.

Bradley Grainger (Logos)

The search results view only shows the first eight hits in each result (in order to save memory); this is by design.

The full results can be seen in the resource:

Bradley Grainger (Logos)

The difference in results 4322 vs 4588 in the same number of verses prompted me check the search results:-

I still get 4322 results.

What is the right result?

Can someone who gets 4,588 results please post a screenshot of the search hit highlighting in the resource? Something with lots of hits, such as Hebrews 8:7-13 might show a difference between our systems. (Or might not.)

Note that it could take 10-20 minutes after running the Bible search before the search hit highlights will appear in the resource.

Bradley Grainger (Logos)

What is the right result?

I programmatically counted all words that are tagged with the ot-quote field in the latest HCSB resource; there are 4,322.

I'm not aware of any indexing bugs that could cause hits to be double-counted, but if there had been a bug in the past, an index built with that bug could be getting its hits merged into any indexes built since then. If anyone still gets 4,588 hits after "rebuild bible index", I'd be very interested to know.

NetworkGeek

https://community.logos.com/discussion/comment/193713#Comment_193713

I got 4588 hits, here is my screen shot of Hebrews 8 so you can see the hits. I just started a reindex of Bible Index, i"ll repost when it is done!

NetworkGeek

https://community.logos.com/discussion/comment/193720#Comment_193720

Here is another graphic showing the Bible itself. My results are bolded for some reason, not highlighted...

Harry Hahne

https://community.logos.com/discussion/comment/193720#Comment_193720

When I do this search on the NA27 Greek New Testament, the matching search terms are not highlighted in the NA27 Bible itself.

Todd Phillips

https://community.logos.com/discussion/comment/193721#Comment_193721

Here is another graphic showing the Bible itself. My results are bolded for some reason, not highlighted...

The bolded text is part of the formatting of the HCSB.

Mark Barnes

https://community.logos.com/discussion/comment/193713#Comment_193713

You need to do some serious optimisation on that search highlighting filter! I ran the search at 20:30, and at 50 minutes I've still no highlighting filter, and one of my cores has been running at 100% all that time. The log file is full of lines like these:

2010-11-18 20:31:12.9055 4 Info TemporaryFileSearchResultReader Creating temporary file for search hits at 'C:\Users\Mark Barnes\AppData\Local\Temp\tmpB9D7.tmp'.
2010-11-18 20:31:13.0355 4 Info TemporaryFileSearchResultReader Wrote 0 hits from 0 results (0.00 MB) in 00:00:00.1278422.

I haven't rebuilt the index yet. I'll do so once I get my CPU back.

<edit> Attached is a PDF of my 4,588 search results if that helps.</edit>

8726.Bible Search.pdf

Mark Barnes

https://community.logos.com/discussion/comment/193734#Comment_193734

The search visual filter finally finished after 69 minutes at 21:39! Results below:

I just tried the search on my other installation (a relatively fresh
installation running 4.1), and got the correct number of results.

Dave Hooton

https://community.logos.com/discussion/comment/193734#Comment_193734

You need to do some serious optimisation on that search highlighting filter! I ran the search at 20:30, and at 50 minutes I've still no highlighting filter, and one of my cores has been running at 100% all that time.

My quad core took at least 10 minutes also running one (of 8) CPU's at near 100%!

Mark Barnes

https://community.logos.com/discussion/comment/193763#Comment_193763

OK, after the reindex I got correct results. Both results in the screenshot below. There are subtle differences (such as the ones at 2:7 and 2:9).

Dave Hooton

https://community.logos.com/discussion/comment/193727#Comment_193727

When I do this search on the NA27 Greek New Testament, the matching search terms are not highlighted in the NA27 Bible itself.

It takes at least 10 minutes depending on your CPU. But I get 4661 results vs your 4548 for NA27, so you should rebuild bible index as Bradley suggests!

NetworkGeek

https://community.logos.com/discussion/comment/193713#Comment_193713

If anyone still gets 4,588 hits after "rebuild bible index", I'd be very interested to know.

I just finished the Rebuild Bible Index, and I still get 4588 results:

Dave Hooton

https://community.logos.com/discussion/comment/193768#Comment_193768

But I get 4661 results vs your 4548 for NA27

Bradley,

I get 44, 442 results in NA27 Int with the same no. of verses as NA27. When I change the search term to greek:* it returns 4661! Why can't it return this result as per the NA27 (why all the extra results)?

Bradley Grainger (Logos)

https://community.logos.com/discussion/comment/193764#Comment_193764

OK, after the reindex I got correct results. Both results in the screenshot below. There are subtle differences (such as the ones at 2:7 and 2:9).

Thanks for taking and posting these screenshots. The search results limits the highlights to 8 hits per verse. 2:7 and 2:9 only have seven highlights, so what is actually happening is that one of those words is counted twice, which is what is inflating the hit counts. I'm still puzzled as to how this is happening.

Bradley Grainger (Logos)

https://community.logos.com/discussion/comment/193769#Comment_193769

I just finished the Rebuild Bible Index, and I still get 4588 results:

The verses that are only showing seven highlights are indicative of a double-counting issue that's inflating your hit counts. If you search for "the" (searching within OT Quotation in HCSB, Match all word forms off) do you get 266 hits in 154 verses, or 532 hits? Does a Basic (not Bible) search for "the" in OT Quotation in HCSB return 266 or 532 results? Do you have version 2009-10-26T21:09:50Z of the resource?

Bradley Grainger (Logos)

https://community.logos.com/discussion/comment/193778#Comment_193778

I get 44, 442 results in NA27 Int with the same no. of verses as NA27. When I change the search term to greek:* it returns 4661! Why can't it return this result as per the NA27 (why all the extra results)?

My guess is that the entire interlinear cell and all its contents are tagged with the ot-quote field, and "*" matches everything.

NetworkGeek

https://community.logos.com/discussion/comment/193784#Comment_193784

The verses that are only showing seven highlights are indicative of a double-counting issue that's inflating your hit counts. If you search for "the" (searching within OT Quotation in HCSB, Match all word forms off) do you get 266 hits in 154 verses, or 532 hits? Does a Basic (not Bible) search for "the" in OT Quotation in HCSB return 266 or 532 results? Do you have version 2009-10-26T21:09:50Z of the resource?

I get 266 results for "the" in 154 verses. HCSB version is 2009-10-26T21:09:50Z .

Not sure completely what's going on or to what degree you have a handle on it, would it help for any of these searches for me to Save as Passage List then export to a file and post so further analysis can be done?

Mark Barnes

https://community.logos.com/discussion/comment/193806#Comment_193806

Dominick,

As you're the only one getting these problems now, how about conducting the search but only in Hebrews 2, and the showing the aligned view? The aligned view might show which words are being double-counted.

Mark

NetworkGeek

https://community.logos.com/discussion/comment/193845#Comment_193845

Dominick,

As you're the only one getting these problems now, how about conducting the search but only in Hebrews 2, and the showing the aligned view? The aligned view might show which words are being double-counted.

Mark

Here ya go - this snapshot shows 4 of the 6 dupes, there were two instances of the word "the" duplicated earlier in Hebrews 2. Thoughts?

NetworkGeek

https://community.logos.com/discussion/comment/193898#Comment_193898

FYI - this problem is on my desktop. I just ran the original search and got the correct number, 4322. I did a resource compare, the two installs have identical resources.

So it seems to be something specific to my one install? Both are also the latest Beta 7.

Should I do a full reindex?

Graham Criddle

https://community.logos.com/discussion/comment/193898#Comment_193898

Here ya go - this snapshot shows 4 of the 6 dupes, there were two instances of the word "the" duplicated earlier in Hebrews 2. Thoughts?

Dominick

I think the first of your "dupes" is actually valid. "Everything" appears at the end of a quotation in verse 8 and the start of another in the same verse.

Graham

NetworkGeek

https://community.logos.com/discussion/comment/193806#Comment_193806

Bradley and Melissa - since no one had any ideas I did a full reindex, and now I got 4322 hits as I should. So the library index file was corrupted some how.

Melissa Snyder

https://community.logos.com/discussion/comment/193991#Comment_193991

Bradley and Melissa - since no one had any ideas I did a full reindex, and now I got 4322 hits as I should. So the library index file was corrupted some how.

Thank you for the update.