Hello Jeremy
Thank you for your advice.
I'll think to store the path as a column.
It seems faster then rcte_tree for search path.
But I'm wondering move operation slower than rcte_tree
So I would like to know much about rcte_tree plugin.
Question
Could you tell me common method to find specified node like ("/a/b/c")
with rcte_tree?
Please tell me if you know alternative approach.
I implemented two method. (find_node_recursive and find_node_ancestors)
Is it possible to use ancestors method for that?
(Is it return sorted node list?)
Thanks
--
Hiroyuki Sato.
P.S.
FYI
rcte_tree plugin does not work OSX marveriks until install sqlite3 >=
3.8.3 by myself.
bundled version is 3.7.13 and it does not support supports common table
expressions.
require 'sqlite3'
require 'sequel'
require 'pp'
def traverse_to_root(node,path)
path << node.name
traverse_to_root(node.parent,path) if node.parent
end
def find_node_recursive(search_path)
last_node = search_path.split('/')[-1]
match_node = nil
FileTree.filter("name = ?",last_node).each do |node|
nodes = []
traverse_to_root(node,nodes)
cur_path = nodes.reverse.join("/")
if( cur_path == search_path )
match_node = node
break
end
end
match_node
end
def find_node_ancestors(search_path)
last_node = search_path.split('/')[-1]
match_node = nil
FileTree.filter("name = ?",last_node).each do |node|
#
# node.ancestors return sorted lists?
#
cur_path = [node.ancestors.reverse.map(&:name),node.name].join("/")
if( cur_path == search_path )
match_node = node
break
end
end
match_node
end
dbpath = File.expand_path(File.dirname(__FILE__),"tree.db")
DB = Sequel.connect("sqlite://#{dbpath}")
unless DB.table_exists?(:file_trees)
# create table
DB.create_table :file_trees do
primary_key :id
String :name
Integer :uid
Integer :gid
foreign_key :parent_id, :file_trees
end
end
class FileTree < Sequel::Model
plugin :rcte_tree, :single_root => true
end
#
# / +- a
# +- b
# +- c
# +- f
# +- d
# +- e
#
unless FileTree.root
FileTree.create(:name => "")
FileTree.create(:name => "a",:parent_id => FileTree.root.id)
b = FileTree.create(:name => "b",:parent_id => FileTree.root.id)
c = FileTree.create(:name => "c",:parent_id => b.id)
FileTree.create(:name => "d",:parent_id => b.id)
FileTree.create(:name => "e",:parent_id => b.id)
f = FileTree.create(:name => "f",:parent_id => c.id)
end
node_f = find_node_recursive("/b/c/f")
pp node_f
node_f = find_node_ancestors("/b/c/f")
pp node_f
----- output
#<FileTree @values={:id=>7, :name=>"f", :uid=>nil, :gid=>nil,
:parent_id=>4}>
#<FileTree @values={:id=>7, :name=>"f", :uid=>nil, :gid=>nil,
:parent_id=>4}>
2014年6月27日金曜日 0時08分24秒 UTC+9 Jeremy Evans:
>
> On Thursday, June 26, 2014 6:22:01 AM UTC-7, Hiroyuki Sato wrote:
>>
>> Dear members.
>>
>> I'm newbie Sequel.
>> Could you tell me best approach and example to represent Unix directory
>> hierarchy with Sequel and Sqlite3?
>>
>> Especially, I would like to know the following
>>
>> How to find '/path/to/deepchild' node with small cost.
>>
>> It's tree, and I found rcte_tree plugin. and I wrote the following code.
>>
>> Alternatively, Should I uses "Nested Sets Model" for that?
>>
>> Sincerely.
>>
>> require 'sqlite3'
>> require 'sequel'
>> require 'pp'
>>
>> dbpath = File.expand_path(File.dirname(__FILE__),"tree.db")
>>
>> DB = Sequel.connect("sqlite://#{dbpath}")
>>
>>
>> unless DB.table_exists?(:file_trees)
>>
>> # create table
>> DB.create_table :file_trees do
>> primary_key :id
>> String :name
>> Integer :uid
>> Integer :gid
>> # And some attributes
>> Numeric :parent_id
>>
>
> Note that using Numeric here is a bad idea. This is a relationship, you
> want to use foreign_key:
>
> foreign_key :parent_id, :file_trees
>
>
>> end
>>
>> end
>>
>> class FileTree < Sequel::Model
>> plugin :rcte_tree, :single_root => true
>> end
>>
>> FileTree.create(:name => "root")
>> FileTree.create(:name => "child",:parent_id => FileTree.root.id)
>>
>> My Goal
>>
>> represent common file operation logging like the following.
>>
>> * create new node in '/path/to/deepchild'
>> * list child in '/path/to/deepchild'
>> * move node from '/path/to/a' to '/new/path/b'
>> * remove node '/path/to/a' and child.
>> * copy node '/path/to/a' to '/new/path/to/b'
>>
>
> Using rcte_tree should work. An alternative approach is just to store the
> path as a column (either instead of or in addition to the name column).
> With an index on the path, most directory operations should be fast,
> probably faster than rcte_tree in some use cases and slower in others.
> Doing this means you don't need to store a foreign key to the parent, you
> just operate on the path. Not sure which way is better.
>
> Thanks,
> Jeremy
>
--
You received this message because you are subscribed to the Google Groups
"sequel-talk" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sequel-talk.
For more options, visit https://groups.google.com/d/optout.
