tFileOutputGML is a Talend Open Studio component writes a file in Graph Modeling Language version 1, compatible with Gephi, Pajek and lots more software for network analysis and social network analysis especially. It’s not a simple file output component, as it supports both nodes and edge writing, talend-to-gml types conversions and some unique features of graph files, like labels and weights. However, it doesn’t support  data structures yet, as same as graphics and icons directives. It still doesn’t validate the graph (ie. edges referring to non-existent or invalid nodes), but this feature is under heavy development.

Download tFileOutputGML for Talend Open Studio

How to use the component

This component support one to two input flow connections. The first is mandatory and it contains nodes definition. You must select the input columns that feed node id and node label respectively.  The second connection is optional, as (technically speaking)  graph are complete, even useless, without edges definition. If you need to adds edge definition in output file, then you must provide a lookup connection describing them, specifying at least source and target id columns. If available, you can activate weight and edge label columns, too. The screenshot on page opening should give you an idea on how this double-connection input element can be used in a real task.

The output schema parameter (the first from upside) let you choose which columns to map as node attributes in output file. Columns are not propagated by default so you can choose which ones to include rather than to drop. Strings are automatically quoted, while lists serialized.

Finally, in advanced parameters you can set some file properties, like character encoding, row separator, graph name and so on.

Available output

Aside writing to a .gml file, the component can feed an optional outgoing main flow connection that propagates the whole set of node attributes that makes the output schema, as set in schema parameter. Please note that no content from lookup connection (edge details) is propagated in this outgoing connection.


  1. aryans

    can i make gml file by taking gimple ssa(pass of gcc) as a input.??????

    • Gabriele

      Probably yes, but sounds really tricky. The component was designed to build a gml file from a de-normalized (first normal form) stream of nodes and edges. For your needs, you would need to pass a named pipe (if you are under Linux) to Talend using tNamedPipe family components, then connect the input stream. Then you would need to parse the tree into a rows+edge pair of flows. Then you would be able to build the file.

      It’s probably much faster to build a Spigot, that’s more or less a importer plugin for Gephy. Here’s some hints:

      Then, you would be able to analyze the network in Gephy itself or export it using Gephy APIs.

      hope this helps!

Submit a Comment

Your email address will not be published. Required fields are marked *

Share This