Bodo 2023.9 Release (Date: 09/01/2023)¶
New Features and Improvements¶
Compilation / Performance improvements:
- BodoSQL generated plans have been further optimized to reduce runtime, compile time, and memory usage.
- Performance and compilation time improvements to several window functions:
LEAD
LAG
AVG
VARIANCE_POP
and equivalent functionsVARIANCE_SAMP
and equivalent functionsSTDDEV_POP
and equivalent functionsSTDDEV_SAMP
and equivalent functionsFIRST_VALUE
LAST_VALUE
RATIO_TO_REPORT
Python coverage:
-
Partial support for the
np.matrix
type with the following functionality (non-distributed):np.asmatrix
to convert a scalar, 1D array, 2D array, list of scalars, or list of 1D arrays to a matrixnp.asarray
to convert a matrix to a 2D array- Addition and subtraction with
+
and-
- Matrix multiplication with
*
,@
ornp.dot
- Calling
len()
on a Matrix - Using
.ndim
,.shape
and.T
(non-distributed) on a Matrix
-
Support for the following Numpy functions:
np.interp
non-distributed (added support for keyword argumentsleft
andright
)np.tile
(added support for specific patterns, see Numpy docs)np.linalg.norm
(added support for keyword argumentaxis=1
when the input is a 2D array)np.nan_to_num
np.dot
(added support for heterogeneous typing between integer & float array inputs)scipy.fftpack.fftshift
(non-distributed)scipy.fftpack.fft2
(non-distributed)
BodoSQL:
- Added support for
HASH(*)
- Added support for
PERCENTILE_CONT
andPERCENTILE_DISC
(non-window support)
2023.9.5 New Features and Improvements¶
Compilation / Performance improvements:
- BodoSQL generated plans have been further optimized to reduce runtime and memory usage.
- Support for executing
UNION
in vectorized mode - Support for executing
ARRAY_AGG
on numeric types in aGROUP BY
2023.9.6 New Features and Improvements¶
Fix critical bugs in vectorized execution mode.
- BodoSQL generated plans have been further optimized to reduce runtime and memory usage.
GET_PATH
and JSON field accesses via:
are supported in some usages.