Period slicer > Exact dates

CorText Manager Q&A forumCategory: Time processingPeriod slicer > Exact dates
tommv asked 4 years ago

Hello dear Cortext Friend,
I am trying to use the period slicer before and after a precise data (i.e. not just the year but the month and the day as well), but I am not sure of the syntax to use to define the exact periods. Would the following be the right format?
ex. [1996-12-13:2016-02-20];[2016-02-21:2016-09-09]

tommv replied 4 years ago

I have also tried to slice my corpus using just the years [1996:2015];[2016:2017], which worked perfectly:

But when I try to run the network analysis with the two custom periods defined above, I only get one map (and no tubes) (

Note that the dynamic network analysis analysis works perfectly when I use the ‘standard periods’ (instead of the ‘custom ones’)
(so it is not a data parsing problem)

I am doing something wrong?

Jean-Philippe Cointet Staff replied 4 years ago

Did you set the correct number of periods (2 I guess) and more importantly regular slicing ?

tommv replied 4 years ago

Well spotted!
I did not set the slicing mode to “regular”.
It works now.

Jean-Philippe Cointet Staff replied 4 years ago

With homogeneous time slicing, the total number of periods is sometimes reduced simply because it is impossible to guarantee an even distribution of documents in each time interval.

1 Answers
Jean-Philippe Cointet Staff answered 4 years ago

Dear Tommv, 
acutally Cortext stores time information as integers only.
It then all depends on the source of your data and the choice you made for processing time information when parsing. It is still possible to produce time periods at the level of days, but you will need to have first transformed all your dates into a simple integer counting the number of days since a fixed date for instance.
This transformation process is automatic for some type of formats (robust csv or xls for instance), otherwise you would need to do the conversion yourself before uploading the file…

tommv replied 4 years ago

Understood. Thanks!