Bodo Error Messages¶
This section lists some of the compilation error messages you may encounter with your jitted functions, reasons for them and suggestions on how to proceed with resolving them.
Unsupported Bodo Functionality¶
BodoError: <functionality> not supported yet
As the error states, this message is encountered when you are attempting to call an as yet unsupported API within a jit function. For example :
would result in an unsupported
Please submit a request for us to support your required functionality here. Also consider joining our community slack, where you can interact directly with fellow Bodo users to find a workaround for your requirements. For longer and more detailed discussions, please join our discourse.
Object Mode can be used to switch to Python interpreted context to be able to run your workload, but we strongly recommend trying to find a Bodo-native workaround.
BodoError: <operation> : <parameter_name> parameter only supports default value
Certain methods only support default parameter values for some of their parameters. Please see supported Pandas API for a list of supported pandas functionality and their respective parameters. We also have a list of supported Numpy , as well as ML operations.
BodoError: <operation>: <operand> must be a compile time constant
Bodo needs certain arguments to be known at compile time to produce an optimized binary. Please see the section on compile time constants for more details.
BodoError: dtype <DataType> cannot be stored in arrays
This error message is encountered when Bodo is unable to assign a supported type to elements of an array.
In this example, we get this error because we attempted to get Bodo to recognize
col1as a column with the datatype
object, and the
objecttype is too generic for Bodo. A workaround for this specific example would be to return
Invalid Series.dt/Series.cat/Series.str, cannot handle conditional yet
This error is encountered when there are conditional assignments of series functions
Series.str, which Bodo cannot handle yet.
When using these operations, you need to include the function and accessor together inside the control flow if it is absolutely necessary. For this specific case, we simply compute the
str.splitwithin the conditional:
Unsupported Numba Errors¶
numba.core.errors.TypingError: Compilation error
This is likely due to unsupported functionality. If you encounter this error, please provide us a minimum reproducer for this error here.
numba.core.errors.TypingError: Unknown attribute <attribute> of type
This is an uncaught error due to unsupported functionality. If you encounter this error, please provide us a minimum reproducer for this error here.