Hi guys,

disclaimer: I did my best to search the forum (and the net..) before starting a new thread, but all I could find were dead links to lists that are not exactly what I am looking for.

I'm looking for an index that lists kanji according to onyomi, and ideally subgroups kanjis of a particular onyomi reading according to radicals/primitives. so, for example, if I was to search for the reading　リン　I would get:
輪倫侖棆淪綸
隣鱗燐麟
林淋琳
etc.

Please note that kanji should be recognized to have more than one onyomi, i.g., the kanji 強 would belong to both the キョウ list and the ゴウ list.

Is there such an index in existence?
Thanks!

This should be easy to make, but how exactly do you want to group "kanjis of a particular onyomi reading according to radicals/primitives"? RTK ordering is easy, anything else will require a specification on database to use.

Grouped by phonetic primitive, sorted by how good the primitive is at predicting on'yomi (I don't remember the exact details of the sorting): http://forum.koohii.com/showthread.php?p...#pid195510

Vempele Wrote:Grouped by phonetic primitive, sorted by how good the primitive is at predicting on'yomi (I don't remember the exact details of the sorting): http://forum.koohii.com/showthread.php?p...#pid195510

Oh my, this is truly frightening in its comprehensiveness.
Thanks a lot!

I wrote a ruby script that uses data from kanjidic and http://cjkdecomp.codeplex.com:

require "nokogiri"

yomi = Hash.new { |x, y| x[y] = [] }
decomp = {}

IO.readlines("cjk-decomp-0.4.0.txt").each { |l|
l =~ /(.+):.+\((.*)\)/
decomp[$1] = $2.split(",")
}

Nokogiri.XML(IO.read("kanjidic2.xml")).css("character").each { |e|
next if e.css("dic_ref[dr_type='heisig']").empty?
kanji = e.css("literal").text
e.css("reading[r_type='ja_on']").each { |x| yomi[x.text] << kanji }
}

yomi.each { |y, kanji|
allcomponents = kanji.map { |x| decomp[x] }.flatten.compact
allcomponents.select { |x| allcomponents.count(x) >= 2 }.uniq.each { |x|
puts y + " " + kanji.select { |k| decomp[k] && decomp[k].include?(x) }.join
}
}

I uploaded the output to http://19a5b0.s3-website-us-west-2.amazo...y-yomi.txt.

The list is limited to RTK kanji, but it includes all yomi in kanjidic and not just jouyou yomi. I didn't try to remove groups of kanji like 綾綸 that share non-phonetic compounds.

Kanji for リン:

リン 倫輪綸
リン 淋琳
リン 燐隣鱗麟
リン 綾綸

Thanks Lauri - that's exactly what I was looking for!

lauri_ranta Wrote:I wrote a ruby script that uses data from kanjidic and http://cjkdecomp.codeplex.com

Oh snap, I've never seen this decomposition dataset before! It looks pretty good, although I did notice a flaw in 留:d(卯,田). Do you have any experience that could let you compare CJKDecomp against CJKVI or KanjiVG, two venerable decomposition databases?