nd.slice does not return empty tensor when begin=end

[lingfanyu]

Description

For mxnet.ndarray.slice(data, begin, end), if begin=end, it does not return an empty tensor. Instead, it returns a tensor with the same shape as data.

Environment info (Required)

----------Python Info----------
Version : 3.5.2
Compiler : GCC 5.4.0 20160609
Build : ('default', 'Nov 23 2017 16:37:01')
Arch : ('64bit', 'ELF')
------------Pip Info-----------
Version : 18.1
Directory : /home/lingfan/.local/pythonenv/dgl/lib/python3.5/site-packages/pip
----------MXNet Info-----------
Version : 1.5.0
Directory : /home/lingfan/.local/pythonenv/dgl/lib/python3.5/site-packages/mxnet
Commit Hash : 812b06a
----------System Info----------
Platform : Linux-4.4.0-1070-aws-x86_64-with-Ubuntu-16.04-xenial
system : Linux
node : ip-172-31-86-85
release : 4.4.0-1070-aws
version : #80-Ubuntu SMP Thu Oct 4 13:56:07 UTC 2018
----------Hardware Info----------
machine : x86_64
processor : x86_64
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 8
On-line CPU(s) list: 0-7
Thread(s) per core: 2
Core(s) per socket: 4
Socket(s): 1
NUMA node(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 79
Model name: Intel(R) Xeon(R) CPU E5-2686 v4 @ 2.30GHz
Stepping: 1
CPU MHz: 2701.332
CPU max MHz: 3000.0000
CPU min MHz: 1200.0000
BogoMIPS: 4600.18
Hypervisor vendor: Xen
Virtualization type: full
L1d cache: 32K
L1i cache: 32K
L2 cache: 256K
L3 cache: 46080K
NUMA node0 CPU(s): 0-7
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single kaiser fsgsbase bmi1 hle avx2 smep bmi2 erms invpcid rtm rdseed adx xsaveopt
----------Network Test----------
Setting timeout: 10
Timing for Conda: https://repo.continuum.io/pkgs/free/, DNS: 0.0039 sec, LOAD: 0.0274 sec.
Timing for Gluon Tutorial(cn): https://zh.gluon.ai, DNS: 0.2937 sec, LOAD: 0.5235 sec.
Timing for PYPI: https://pypi.python.org/pypi/pip, DNS: 0.0060 sec, LOAD: 0.0966 sec.
Timing for FashionMNIST: https://apache-mxnet.s3-accelerate.dualstack.amazonaws.com/gluon/dataset/fashion-mnist/train-labels-idx1-ubyte.gz, DNS: 0.0047 sec, LOAD: 0.1268 sec.
Timing for MXNet: https://github.com/apache/incubator-mxnet, DNS: 0.0015 sec, LOAD: 0.4607 sec.
Timing for Gluon Tutorial(en): http://gluon.mxnet.io, DNS: 0.0451 sec, LOAD: 0.3926 sec.

Minimum reproducible example

>>> import mxnet.ndarray as nd
>>> a = nd.normal(shape=(4, 3))
>>> a
[[ 2.2122064   0.7740038   1.0434405 ]
 [ 1.1839255   1.8917114  -1.2347414 ]
 [-1.771029   -0.45138445  0.57938355]
 [-1.856082   -1.9768796  -0.20801921]]
<NDArray 4x3 @cpu(0)>
>>> nd.slice(a, begin=0, end=0)
[[ 2.2122064   0.7740038   1.0434405 ]
 [ 1.1839255   1.8917114  -1.2347414 ]
 [-1.771029   -0.45138445  0.57938355]
 [-1.856082   -1.9768796  -0.20801921]]
<NDArray 4x3 @cpu(0)>
>>> nd.slice(a, begin=2, end=2)
[[-1.7710290e+00 -4.5138445e-01  5.7938355e-01]
 [-1.8560820e+00 -1.9768796e+00 -2.0801921e-01]
 [ 0.0000000e+00  0.0000000e+00  1.8637270e-43]
 [ 0.0000000e+00  3.8537848e-34  4.5786026e-41]]
<NDArray 4x3 @cpu(0)>

@zheng-da @jermainewang

[leleamol]

Thank you for submitting the issue! I will try and suggest some labels so that the appropriate MXNet community members can help resolve it.

[leleamol]

@mxnet-label-bot add [Python, NDArray, Operator, Bug]

[wkcn]

In src/operator/tensor/matrix_op-inl.h, there is no judgement for e == b in the function SetSliceOpOutputDimSize.

[szha]

Empty tensor is currently not possible to be explicitly expressed in MXNet.

[anirudh2290]

Since the begin=end case is currently returning ndarray with garbage values, should we throw an exception when we encounter these inputs. I know we may have users depending on this behavior, but the number may be very less because of the garbage values.