Unnecessary variants for single address

[ghost]

grep -a "愛知県名古屋市南区豊田町" mecab-user-dict-seed.20160225.csv

名古屋市豊田町,1293,1293,-5820,名詞,固有名詞,地域,一般,,,愛知県名古屋市南区豊田町,ナゴヤシトヨダチョウ,ナゴヤシトヨダチョー
愛知県南区豊田町,1293,1293,-1981,名詞,固有名詞,地域,一般,,,愛知県名古屋市南区豊田町,アイチケンミナミクトヨダチョウ,アイチケンミナミクトヨダチョー
愛知県名古屋市南区豊田町,1293,1293,-19354,名詞,固有名詞,地域,一般,,,愛知県名古屋市南区豊田町,アイチケンナゴヤシミナミクトヨダチョウ,アイチケンナゴヤシミナミクトヨダチョー
愛知県名古屋市豊田町,1293,1293,-18608,名詞,固有名詞,地域,一般,,,愛知県名古屋市南区豊田町,アイチケンナゴヤシトヨダチョウ,アイチケンナゴヤシトヨダチョー

I think we don't need "名古屋市豊田町" "愛知県南区豊田町" "愛知県名古屋市豊田町".
https://www.google.co.jp/search?q="名古屋市豊田町"
4 results
https://www.google.co.jp/search?q="愛知県南区豊田町"
0 results
https://www.google.co.jp/search?q="愛知県名古屋市豊田町"
0 results

[neologd]

Thank you for your comment.

When we use Google Custom Search API, we should pay 5$/1000 queries.
We can't use that filtering method.

And we think these entries are important to aggregate an unknown address string.
We need these variants.

In the future, we will provide an option for people who want to save RAM.
I think that your opinion will be reflected in at that time.

[ghost]

I thought you generated the entries automatically from
Japan Post's "KEN_ALL.CSV".

e.g.
23112,"457 ","4570855","アイチケン","ナゴヤシミナミク","トヨダチョウ", "愛知県","名古屋市南区","豊田町",0,0,0,0,0,0

Ken + Chou ("愛知県" + "豊田町") is unnecessary in most cases,
so I thought you could remove the code that generates "Ken + Chou".

we think these entries are important to aggregate an unknown address string.

hm.
We don't see "Ken + Chou" on the internet in most cases,
and I think splitted "Ken|Chou" is enough.
Anyway this is not a bug and this is my opinion.

Thank you for the reply.