Skip to content

Workaround for ECMWF GRIB API bug. #6

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

Workaround for ECMWF GRIB API bug. #6

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 22 additions & 4 deletions lib/iris/fileformats/grib/_save_rules.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -298,12 +298,30 @@ def grid_definition_template_12(cube, grib):
gribapi.grib_set(grib, "scaleFactorAtReferencePoint",
cs.scale_factor_at_central_meridian)

# Check that scaleFactorAtReferencePoint is being stored correctly.
# Check for a bug in the ECMWF GRIB API (at version 1.12.1) where
# the scale factor is erroneously classified as a signed 32-bit
# integer.
scale_at_ref_point = gribapi.grib_get(grib, "scaleFactorAtReferencePoint")
if cs.scale_factor_at_central_meridian != scale_at_ref_point:
msg = "GRIBAPI error prevented correct setting of "\
"key 'scaleFactorAtReferencePoint'."
raise iris.exceptions.TranslationError(msg)
# As a workaround, we can call grib_set with an integer value
# whose *on-disk* bit pattern corresponds to the desired bit
# pattern of the floating-point value.
scale_as_float32 = np.array(cs.scale_factor_at_central_meridian,
dtype='f4')
scale_as_uint32 = scale_as_float32.view(dtype='u4')
if scale_as_uint32 >= 0x80000000:
# Convert from two's-complement to sign-and-magnitude.
# NB. Because of the silly representation of negative
# integers in GRIB2, there is no value we can pass to
# grib_set that will result in the bit pattern 0x80000000.
# But since that bit# pattern corresponds to a floating
Copy link
Owner

@DPeterK DPeterK Dec 2, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is "bit#" deliberate (as in "bit-number"), or a typo?

Copy link
Author

@rhattersley rhattersley Dec 2, 2014

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Typo! 😒

# point value of negative-zero, we can safely treat it as
# positive-zero instead.
scale_as_grib_int = 0x80000000 - int(scale_as_uint32)
else:
scale_as_grib_int = int(scale_as_uint32)
gribapi.grib_set(grib, 'scaleFactorAtReferencePoint',
scale_as_grib_int)


def grid_definition_section(cube, grib):
Expand Down