descriptive statistic - weekend effect

[Marketa]

Hello,
I would like to ask you if is there any possibility to import data with different long time series. I tried to analyse a weekend effect for monday - friday but each year has different number of data. For examle in the year 2012 monday has (47) days, tuesday (50), wednesday (51), thursday (51), friday (51). How to import data to EViews 7? It´s to hard to analyse each day of week for each year.
Thank you for your reply
Marketa

[EViews Gareth]

I'm confused. What structure does your data have? Why is it difficult to bring into EViews?

[Marketa]

let´s check example bellow:

Code: Select all

date monday date tuesday date wednesday date thursday date friday 4.1.2010 1,954071699 5.1.2010 0,383200949 6.1.2010 0,309925808 7.1.2010 0,219745641 8.1.2010 0,424042829 11.1.2010 0,318668586 12.1.2010 -1,06077448 13.1.2010 0,80659384 14.1.2010 0,247065386 15.1.2010 -1,239165864 25.1.2010 0,60296246 19.1.2010 1,174165392 20.1.2010 -1,541302695 21.1.2010 -2,142482578 22.1.2010 -2,025402445 1.2.2010 1,791725306 26.1.2010 -0,635539619 27.1.2010 0,103669467 28.1.2010 -1,123748658 29.1.2010 -1,057898986 8.2.2010 -1,020708859 2.2.2010 1,321240765 3.2.2010 -0,831732036 4.2.2010 -3,68445375 22.2.2010 -0,06665829 11.2.2010 1,160545588

when i want to put it together into Eviews I have to specify number in each column - see attachments (number of observations) - it differs for each day
I suppose that it will be undated data series because each day of the week has different date
So am I right? Is this process by undated data series?
It is not working when I state number of observations as 6 (because of the thurday and monday), others days of the week has smaller number of dates.
The issue is that I have to put each day for each year separately and it takes long time.
Due to this fact I tried to put all days together for each year and test it together for descriptive statistic.

Thank you for your reply

[EViews Gareth]

Could you post the data file? This should be easy...

[Marketa]

of course :)

the file is too large for direct upload to this forum. So I put it to Dropbox, see this link:
https://dl.dropboxusercontent.com/u/142 ... 202010.xls

Thank you

[EViews Gareth]

Ha, that's one of the craziest data formats I've seen :)

There are lots of ways to do it. I think this is the easiest (in EV7 or .

Code: Select all

wfcreate d5 2010 2010 import(mode=ms) "nya 2010.xls" range="a1:b53" @id date @destid @date @rename monday x import(mode=ms) "nya 2010.xls" range="c1:d53" @id date @destid @date @rename tuesday x import(mode=ms) "nya 2010.xls" range="e1:f53" @id date @destid @date @rename wednesday x import(mode=ms) "nya 2010.xls" range="g1:h53" @id date @destid @date @rename thursday x import(mode=ms) "nya 2010.xls" range="i1:j53" @id date @destid @date @rename friday x

[Marketa]

this information is not useful for me. The reason is that I want to have 5 columns - for each day descriptive statistic separately. Not all data in one column together :(

[EViews Gareth]

That's sort of weird, but, if you insist that's what you want, simply lop off the "@rename" bit at the end of each import command.