giraph-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mckie, Duncan" <dmc...@ebay.com>
Subject Re: Need help on simple text based giraph input format
Date Tue, 06 Jan 2015 16:04:03 GMT
Hi Arghya,

I'm not sure if you received a reply or not, but here is a Text-based Vertex Adjacency List
that I created:

https://github.com/duncanmckie/giraph/blob/release-1.1/giraph-examples/src/main/java/org/apache/giraph/examples/io/formats/TextTextNullTextInputFormat.java

This takes a tab-separated, vertex-oriented adjacency list as an input, and (as per your example)
does not expect any edge weights.

You can easily create your own input format by customising the examples provided, changing
the Writable classes and doing any preprocessing as required.

Cheers,

Duncan


From

Arghya Kusum Das <arghyakusumdas2...@gmail.com>

Subject

Need help on simple text based giraph input format

Date

Fri, 05 Dec 2014 01:36:14 GMT

Hi,
This is the first time I am trying to deal with giraph input format class
and need some help.

I want a input-format-class where vertex-id, vertex-value and edge-weight
everything will be simple text.

Eg. I have a simple Graph like following:

AAA  AAT AAG
AAC  ACG ACT
AAG AGT AGA
AAT ATT
ACG CGA
ACT CTG CTT CTC

The first column is the vertex id and the following are the edges
(Adjacency list). As you can see edge weight or vertex id can be null also.

Is there any predefined input-format in giraph for this?
If so what is that? And if not can anybody provide a simple class for that?

--> It is not possible to come up with different integer ids for different
nodes. Because the vertex id can be even more than 50 characters (It is
just a simple excerpt).  So I need a very simple text based vertex input
format where each line represents a vertex and its outgoing edges
(adjacency list).

--
Thanks and regards,
Arghya Kusum Das
(225-270-6163)



Mime
View raw message