More about Sitecatalyst character encoding issues
I’ve been mentioning character encoding problems before, but during the last week I discovered even more of these problems, so it’s time for another post on the subject.
Under certain circumstances the reports from SiteCatalyst can become really … well lets just say, less useful.
It may be difficult to imagine how a few character encoding issues can affect SiteCatalyst that much, a demonstration is needed.
On our site we have a search enabled FAQ, we capture all search phrases in a prop, this prop is correlated with another prop containing information about our Site Sections, we use this to break down the search phrases into Business and Home user search phrases.
A report could look like this: ( The date showing 1 Jan 1969 – 1 Jan 1970 is just a sideeffect of the latest Omniture maintenance release, and not relevant for this demonstration)
Now let’s try to apply a filter containing some Danish words… Like: netværk OR spærring
In return we get a report with no results … strange, after some double checking it’s safe to conclude that filters containing Danish characters won’t work when a correlation has been applied as well.
Let’s try removing the Danish characters and repeat, so now we filter : netv OR rring
Oh yes, much better, but the graph isn’t nice looking anymore, it contains some strange mix of Danish and I guess it’s Japanese.
Now what? such a report isn’t really suitable for distribution. I guess we could download a PDF version instead? Bad idea, that’s even more broken. Graph is the same, but text gets truncated at first occurrence of a Danish character.
Excel or Word downloads ? Well unfortunately they are also broken.
