I'd like to have 117.3(11.1) parse as 117.3 \pm 11.1. Is this possible to achieve in a table? Currently sunitx somehow regards the bracketed value as too small so it's all wrong.

Note: this question about parsing numbers in a table, not via \num in running text, such as addressed in this question as of 4 days ago.

version of siunitx

Package: siunitx 2014/09/27 v2.6b A comprehensive (SI) units package

Picture of MWE

enter image description here






    \item Status quo: \verb+\num{111.26(13.45)}+: \num{111.26(13)} (obviously I omitted the \verb+.45+ as it wouldn't work)
    \item What I want: to have \verb+\num{111.26(13.45)}+ parse as \num{111.26 \pm 13.45}
{Status quo} & {Wish}\\
111.26(13) & \num{111.26 \pm 13.45}\\
  • Is there a reason why you're not inputting the material as 111.26(1345) in the table and as \num{111.26(1345)} in running text? – Mico Oct 7 '14 at 8:20
  • @Mico The data comes straight out of Excel by the help of Excel2LateX and so there it is. Why would anyone delete the decimal markers in the original data, whether it's Excel, Calc, some .csv-file...? – henry Oct 7 '14 at 8:23
  • That's a question for the package's author and maintainer, I suppose. – Mico Oct 7 '14 at 8:26
  • 2
    Short answer here is going to be 'no' as 111.26(13.45) is wrong: bracket uncertainties refer to the least significant digits. Here your accuracy also does not support the number of figures you have: 111(13) is all you can say. – Joseph Wright Oct 7 '14 at 8:30
  • 1
    @henry To be clear, when I say 'wrong' I mean _siunitx uses a defined format for uncertainties and it does not cover the format you use'. The code already has to deal with a variety of input formats, and adding more would be very tricky! (I am working on version 3 and will consider this, but it looks quite hard.) – Joseph Wright Oct 7 '14 at 8:40

Just as you would write \num{111.26(1345)} in running text, you should write 111.26(1345) in an S column.

Aside: As Joseph Wright has already observed in a comment, you're providing an unrealistic amount of precision with regard to the uncertainty component; 111(13) would be more useful and realistic.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.