Archivos Mensuales: octubre 2015

An Eviews workfile reader written in Python

March 20h, 2016 update: the Eviews workfiles reader described below also works with Eviews9 workfiles.

On a side note …

In the video below, I show in action the Eviews workfile (*.wf1) reader I wrote in Python.

There is almost no information on the Eviews workfiles format (it is proprietary after all…). The only resource I found out on the internet is located here: EViews workfile format . But, I cannot say it was actually helpful. My understanding of how Eviews files are structured and the way data can (must?) then be extracted (the ‘algorithm’) are very different, to be honest. This probably relates to Eviews versions. The workfiles I worked on were created with Eviews 8.

For the time being, the Python class I wrote returns all the time series as a single pandas.DataFrame object. Alternatively, it could return a specific series (or set of series) depending on user’s needs. It is straightforward to implement.

There are pieces of information contained in Eviews workfiles that are left to be ‘decoded’, eg dates when objects were last updated. It is something I have even not tried to do. All I was willing here was exporting the time series/observations from  Eviews to Python.

Et voilà!