[openspending-dev] [OpenSpending] incorrect csv format?
friedrich.lindenberg at okfn.org
Tue Sep 25 21:51:42 UTC 2012
Argh, I was being fuzzy - you can of course use all unicode in the
data itself, it's just column headers that are weird. Other than that,
this looks very good!
On Tue, Sep 25, 2012 at 11:36 PM, Christian Sawatzki
<sawatzki at osnanet.de> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> Am 25.09.2012 22:42, schrieb Friedrich Lindenberg:
>>>> Moving to openspending-dev for technobabble :)
>> Hi Christian!
>> two comments: I think you have a lot of columns with umlauts, one
>> of them even seems to have a line break. OS currently has some
>> issues with this, perhaps you could normalize things down to
>> ae/ue/oe with underscores? Also, while you can load the data like
>> this, you may want to transpose it first: Make "Bereich" the first
>> level of your dataset and then establish a new column, say
>> "Economic Type", with the current column names as values and the
>> actual number in a third column. This will also allow you to
>> aggregate the numbers to get overall revenue for a Bereich. Also,
>> as a matter of convention, OS requires each dataset to have a
>> temporal column. If you don't have one, just add a fake one with
>> "2012" in all rows.
>> Hope this helps,
>> - Friedrich
> Hello Friedrich,
> thank you for the help. You mean we should design the file like this?
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.11 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://www.enigmail.net/
> -----END PGP SIGNATURE-----
More information about the openspending-dev