No observation error - all data filtered

[Jenna Grabey]

My problem with Stata 15.0 concerns the r(2000) error: no observations. You have requested some statistical calculation and there are no observations on which to perform it. Perhaps you specified if or in and inadvertently filtered all the data.

This makes sense because when I manually check my data I can see that my code has indeed filtered all of the data. However as I am using Stata to clean data and check for errors this is actually a good thing (it means the data that piece of code was checking is clean and free of errors). So how do I run this code but make it so that if the code does filter all the data it just carries on to the next chunk of code?

The code I am using is:

putdocx paragraph
putdocx text ("ILI 3 date of onset is not between the baseline date and the 30th April 2018")
putdocx table TRIALNAME = data(StudySubjectID VISDAT_E1_C1 ILISTDAT_E7_C17_3) if ///
ILISTDAT_E7_C17_3 > td(30apr2018) & ILISTDAT_E7_C17_3 !=.

I have also tried using:

if !_rc {
putdocx paragraph
putdocx text ("ILI 3 date of onset is not between the baseline date and the 30th April 2018")
putdocx table TRIALNAME = data(StudySubjectID VISDAT_E1_C1 ILISTDAT_E7_C17_3) if ///
ILISTDAT_E7_C17_3 > td(30apr2018) & ILISTDAT_E7_C17_3 !=.
}

The latter prevented the error from occurring, but when I entered dummy data of a date that the code should pull out (05/05/2018) it didn’t pull this out.

Example data (this is made up data for confidentiality reasons but this reflects the type of data I am working with) (I read about dataex but it was not clear from the guidance on how to implement this in a forum pot, I hope what I have done below is satisfactory, if not the please let me know how to amend what I have done and I will do so):

 
StudySubjectID
VISDAT_E1_C1
ILISTDAT_E7_C17_3

100
06/10/2017
15/02/2018

101
06/10/2017


102
10/10/2017
22/02/2018

103
21/10/2017
03/03/2018

104
23/10/2017
12/12/2017

105
30/10/2017


106
30/10/2017


107
31/10/2017
05/01/2018

108
01/11/2017


109
01/11/2017


110
02/11/2017
06/04/2018

110
02/11/2017
23/03/2018

110
02/11/2017
21/12/2017

110
02/11/2017


110
02/11/2017


110
02/11/2017


110
03/11/2017
15/01/2018

110
03/11/2017


110
03/11/2017


110
04/11/2017
29/04/2018

Thank you in advance for any advice.

Best wishes,
Jenna

[Nick Cox]

You show three variables there. You need to show the results of

describe  StudySubjectID    VISDAT_E1_C1    ILISTDAT_E7_C17_3

so that we can see storage types and display formats. (That is what dataex does.)

I've never used putdocx so I may not reply, but I think people who do use it need this information. My wild guess is that your dates are still string variables.

[Jenna Grabey]

You show three variables there. You need to show the results of

describe StudySubjectID VISDAT_E1_C1 ILISTDAT_E7_C17_3

so that we can see storage types and display formats. (That is what dataex does.)

I've never used putdocx so I may not reply, but I think people who do use it need this information. My wild guess is that your dates are still string variables.

Thank you for your advice Nick. I have checked at the dates are in integer format:


StudySubjectID str7 %9s
VISDAT_E1_C1 int %td
ILISTDAT_E7_C~3 int %td

[Nick Cox]

OK. So I note that in your example the dates after 30 April 2018 do not occur for the third variable.

[Jenna Grabey]

Correct, which means the data is in fact clean. So ideally Stata would just skip this and move on rather than sending out an error message. As I am working with a large dataset I'd like to be able to clean that data using a Stata program rather than manually checking the data.

[Jenna Grabey]

I managed to fix this issue by using "capture", see amended code:

capture quietly {
putdocx paragraph
putdocx text ("ILI 3 date of onset is not between the baseline date and the 30th April 2018")
putdocx table TRIALNAME = data(StudySubjectID VISDAT_E1_C1 ILISTDAT_E7_C17_3) if ///
ILISTDAT_E7_C17_3 > td(30apr2018) & ILISTDAT_E7_C17_3 !=.
}

Hope this can help someone else out too!