eppy converts numbers in scientific notation to plain decimal notation, which might be out of range when processed by E+

[wannizhang]

It seems that when parsing numbers, eppy will convert those scientific notations (such as 1e+10) to plain decimal notations. I got a idf file with an object like this:

  ScheduleTypeLimits,
    AnyValue,                !- Name
    -1e+018,                 !- Lower Limit Value
    1e+019,                  !- Upper Limit Value
    Continuous;              !- Numeric Type

After parsed by eppy, it becomes

ScheduleTypeLimits,
    AnyValue,                 !- Name
    -1000000000000000000,     !- Lower Limit Value
    10000000000000000000,     !- Upper Limit Value
    Continuous;               !- Numeric Type

Then when it's simulated with E+, an error ** Severe ** stoll: out of range will occur, because the numbers have too many digits.
Not sure if this is a problem easy to fix in the source code. Otherwise I can always check the idf files to remove those extreme large numbers.
Thanks in advance!

[santoshphilip]

I see this as a critical bug.
It has to be fixed.

I need a couple of days to get to this.
And I might need some more input from you.

Can you tell me which versions of E+ this happened in.
I don't think the E+ version number is significant, But it is worth looking into, to see if anything in the IDD file changed

[wannizhang]

Thank you! I'm using E+ 9.2. FYI, it happens both on Windows and Mac.

[santoshphilip]

@wannizhang , Can you paste the relevant parts of the error message here
Is it possible to figure out how many digits is E+ capable of reading looking at the error message ?

[santoshphilip]

re-reading your first comment, It is clear that the limitation is E+, not in eppy.
to be practical, eppy should be coded with a workaround for that limitation.

[wannizhang]

@santoshphilip Yes, I think it's a limit of E+. The .err file of E+ doesn't show the specific limit, but I tried that it allows 20 digits at most (1e19 / -1e18).
Since using scientific notations won't be out of range, I wonder if eppy can keep the scientific notations in the original file when parsing the idf.

[santoshphilip]

• I can keep scientific notation
• I don't want it to be scientific notation everywhere. Makes it hard to read.
• If it is greater than x digits, I'll let it write in scientific notation.

Can you test less number of digits. Try 19, 18, 17, 16, 15. Find out where E+ stops failing. If you do that, I can update the code with some level of confidence that we are doing the right thing.

To clarify what is happening internally:
Eppy is seeing it as a floating point number or as an integer. Once eppy reads the file, it has no format (scientific notation is a printing format). it is just a number. It becomes a scientific notation or some other format, only when it saved to the disk or printed to screen.

[wannizhang]

@santoshphilip Yeah I've tried different numbers of digits and it starts to fail at 20 digits (19 works).
Sure. It makes sense to only keep the scientific notation when the number is out of range.
Thanks!

[santoshphilip]

I fixed this in branch i289_scientific
I put the cutoff at 18 digits (put a little safety margin there)

@wannizhang , can you do an E+ run and test
If it looks good I'll merge into develop branch

[santoshphilip]

merged into develop