search

himselfv / wakan

Japanese and Chinese learning tool with dictionary
38 stars 7 forks source link

Some kanji not shown in KanjiList when ordering by Frequency #279

Closed himselfv closed 9 years ago

himselfv commented 9 years ago

Original report by me.

Originally reported on Google Code with ID 279

In latest builds (Wakan 1.90) some kanji are not shown in KanjiList when it is ordered
by Frequency, even if no filters are enabled.

This is evident if you switch between Frequency and e.g. Stroke Count ordering on a
small kanji set. Frequency ordered set is usually much smaller than Stroke Count ordered
one.

Examples of kanji that are not shown: 噌, 凰.

The reason was that KanjiList expects JpFrequency_Ind index to be also ordered by "Japanese
only"/"Chinese too" flag when in Japanese mode. It scans until the first "Chinese"
character and ignores the rest.

While only "Japanese" characters get assigned "Japanese frequency" and sorted to the
top, some "Japanese" characters lack frequency info. They were sorted together with
Chinese ones, in no specific order.

Fixed by adding Chinese+ column to the index. Now the index is initially sorted by
Chinese/Japanese flag, and then by Frequency. (The opposite would work the same: sort
by frequency, and then by flag)

Fixed in rev. 5e0e0c9a623f.

Reported by himselfv on 2015-04-06 12:24:11

himselfv commented 9 years ago

Original comment by Anonymous.

Hi, I can't seem to confirm that this issue has been solved.

On any ordering (with all components up to date) it shows me Found Kanji (2762), but
on Freq ordering it shows (2366) This is in Japanese mode with "common only".

Without any filter it's (6355) vs. (2501)

We are talking about this, right? So reopening for now.

I have to say that the one major thing I changed in the settings was to remove many
entries from Kanji Details, basically leaving only radicals, on-kun and Nanori.

Reported by supermarkus420 on 2015-08-09 09:39:58

himselfv commented 9 years ago 
Please try rebuilding the character database:
1. Database -> Download/update components, check kanjidic and unihan and proceed.
2. Database -> Settings -> Database maintenance -> Import character data -> OK.

The problem was in the character database so without rebuilding it won't go away (I'll
put the updated db with the next release).

Reported by himselfv on 2015-08-10 12:36:07

himselfv commented 9 years ago

Original comment by Anonymous.

Confirmed. After doing step 2 it now shows all the Kanji, as expected. Nice to know
that I now have up to date Kanji information. I have been yearning for this since when
I found out about Wakan and the Joyou Kanji actually got reformed back in 2011, was
it?

This is very nice. Thx.

Reported by supermarkus420 on 2015-08-11 15:34:41