Warning! ***HDF5 library version mismatched error*** python pandas windows
Question:
I’m using pandas/python to save a DataFrame in a HDFStore format. When I apply the my_data_frame.to_hdf(arguments…) command I have an error message:Warning! ***HDF5 library version mismatched error *** and my program is stopped.
I’m working on Windows 7 (64bits), using Python 3.5.2 :: Anaconda 4.1.1 (64-bit).
I’ve been reading about this error message and as it says it’s a problem between the version of HDF5 installed on my computer and the one used by Anacondas. According this post, a simple “conda install -c anaconda hdf5=1.8.18” could resolve my problem but I’m still having the same message error.
Thanks for your help guys.
Here I put a complete log of the error:
Warning! ***HDF5 library version mismatched error***
The HDF5 header files used to compile this application do not match
the version used by the HDF5 library to which this application is linked.
Data corruption or segmentation faults may occur if the application continues.
This can happen when an application was compiled by one version of HDF5 but
linked with a different version of static or shared HDF5 library.
You should recompile the application or check your shared library related
settings such as 'LD_LIBRARY_PATH'.
You can, at your own risk, disable this warning by setting the environment
variable 'HDF5_DISABLE_VERSION_CHECK' to a value of '1'.
Setting it to 2 or higher will suppress the warning messages totally.
Headers are 1.8.15, library is 1.8.18
SUMMARY OF THE HDF5 CONFIGURATION
=================================
General Information:
-------------------
HDF5 Version: 1.8.18
Configured on: 2017-05-31
Configured by: NMake Makefiles
Configure mode: CMAKE 3.8.0
Host system: Windows-6.3.9600
Uname information: Windows
Byte sex: little-endian
Libraries:
Installation point: C:/bld/hdf5_1496269860661/_b_env/Library
Compiling Options:
------------------
Compilation Mode: RELEASE
C Compiler: C:/Program Files (x86)/Microsoft Visual Studio
14.0/VC/bin/amd64/cl.exe
CFLAGS: /DWIN32 /D_WINDOWS /W3
H5_CFLAGS:
AM_CFLAGS:
CPPFLAGS:
H5_CPPFLAGS:
AM_CPPFLAGS:
Shared C Library: YES
Static C Library: YES
Statically Linked Executables: OFF
LDFLAGS: /machine:x64
AM_LDFLAGS:
Extra libraries: C:/bld/hdf5_1496269860661/_b_env/Library/lib/z.
lib
Archiver:
Ranlib:
Debugged Packages:
API Tracing: OFF
Languages:
----------
Fortran: OFF
Fortran Compiler:
Fortran 2003 Compiler:
Fortran Flags:
H5 Fortran Flags:
AM Fortran Flags:
Shared Fortran Library: YES
Static Fortran Library: YES
C++: ON
C++ Compiler: C:/Program Files (x86)/Microsoft Visual Studio
14.0/VC/bin/amd64/cl.exe
C++ Flags: /DWIN32 /D_WINDOWS /W3 /GR /EHsc
H5 C++ Flags:
AM C++ Flags:
Shared C++ Library: YES
Static C++ Library: YES
Features:
---------
Parallel HDF5: OFF
High Level library: ON
Threadsafety: ON
Default API Mapping: v18
With Deprecated Public Symbols: ON
I/O filters (external): DEFLATE
MPE:
Direct VFD:
dmalloc:
Clear file buffers before write: ON
Using memory checker: OFF
Function Stack Tracing: OFF
Strict File Format Checks: OFF
Optimization Instrumentation:
Answers:
Headers are 1.8.15, library is 1.8.18
Your error information shows, then, that you need install the 1.8.15
version.
conda install -c anaconda hdf5=1.8.15
In case someone is having this problem and no amount of uninstalling with pip
or conda
or anything will work: I went through every /lib
path and manually deleted libhdf5*
files, then installed hdf5
with macports
.
Try to uninstall h5py module and install it again. This fixed it for me
pip uninstall h5py
pip install h5py
The only procedure that worked for me on macOS was to create a virtualenv:
virtualenv -p python3 myenv
. myenv/bin/activate
pip3 install h5py==1.10.4 # or whichever version you want
Absolutely nothing else worked!
You may need a brew install hdf5
github issue that did not help:
You can force a custom install of a specific package version:
conda install --force-reinstall anaconda hdf5==1.8.15
or
conda install -c conda-forge hdf5=1.8.15
For some reason uninstalling with conda and then installing hdf5 again with conda is not working. However, if you use pip to uninstall and then conda to install hdf5, it’s working.
So try:
pip uninstall hdf5
conda install hdf5
I tried every method here but only something else entirely worked –
conda install -c conda-forge hdf5=1.10.5 # newer system version
conda install -c conda-forge hdf5=1.8.18 # for this particular problem
This will forcefully install – by either upgrading or downgrading – your hdf5. This in turn will uninstall keras with it and when you reinstall keras that upgrades the hdf5.
My case was trivial and did not need any installations: I had started spyder
and an IPython
console from a conda
environment using cmd
:
activate env
spyder
I then upgraded hdf5
for the env
in cmd
using conda
. However, after running my code in the same IPython
console, I got a similar error message. All it took was to restart spyder
and the corresponding IPython
console. I suppose this reloads the packages consistently.
For people not using conda. As my hdf5 library on Debian was not as new as pip3 supposed, it helped me to install the python3-h5py package using apt as a root, instead of doing it locally via pip3.
I’m using pandas/python to save a DataFrame in a HDFStore format. When I apply the my_data_frame.to_hdf(arguments…) command I have an error message:Warning! ***HDF5 library version mismatched error *** and my program is stopped.
I’m working on Windows 7 (64bits), using Python 3.5.2 :: Anaconda 4.1.1 (64-bit).
I’ve been reading about this error message and as it says it’s a problem between the version of HDF5 installed on my computer and the one used by Anacondas. According this post, a simple “conda install -c anaconda hdf5=1.8.18” could resolve my problem but I’m still having the same message error.
Thanks for your help guys.
Here I put a complete log of the error:
Warning! ***HDF5 library version mismatched error*** The HDF5 header files used to compile this application do not match the version used by the HDF5 library to which this application is linked. Data corruption or segmentation faults may occur if the application continues. This can happen when an application was compiled by one version of HDF5 but linked with a different version of static or shared HDF5 library. You should recompile the application or check your shared library related settings such as 'LD_LIBRARY_PATH'. You can, at your own risk, disable this warning by setting the environment variable 'HDF5_DISABLE_VERSION_CHECK' to a value of '1'. Setting it to 2 or higher will suppress the warning messages totally. Headers are 1.8.15, library is 1.8.18 SUMMARY OF THE HDF5 CONFIGURATION ================================= General Information: ------------------- HDF5 Version: 1.8.18 Configured on: 2017-05-31 Configured by: NMake Makefiles Configure mode: CMAKE 3.8.0 Host system: Windows-6.3.9600 Uname information: Windows Byte sex: little-endian Libraries: Installation point: C:/bld/hdf5_1496269860661/_b_env/Library Compiling Options: ------------------ Compilation Mode: RELEASE C Compiler: C:/Program Files (x86)/Microsoft Visual Studio 14.0/VC/bin/amd64/cl.exe CFLAGS: /DWIN32 /D_WINDOWS /W3 H5_CFLAGS: AM_CFLAGS: CPPFLAGS: H5_CPPFLAGS: AM_CPPFLAGS: Shared C Library: YES Static C Library: YES Statically Linked Executables: OFF LDFLAGS: /machine:x64 AM_LDFLAGS: Extra libraries: C:/bld/hdf5_1496269860661/_b_env/Library/lib/z. lib Archiver: Ranlib: Debugged Packages: API Tracing: OFF Languages: ---------- Fortran: OFF Fortran Compiler: Fortran 2003 Compiler: Fortran Flags: H5 Fortran Flags: AM Fortran Flags: Shared Fortran Library: YES Static Fortran Library: YES C++: ON C++ Compiler: C:/Program Files (x86)/Microsoft Visual Studio 14.0/VC/bin/amd64/cl.exe C++ Flags: /DWIN32 /D_WINDOWS /W3 /GR /EHsc H5 C++ Flags: AM C++ Flags: Shared C++ Library: YES Static C++ Library: YES Features: --------- Parallel HDF5: OFF High Level library: ON Threadsafety: ON Default API Mapping: v18 With Deprecated Public Symbols: ON I/O filters (external): DEFLATE MPE: Direct VFD: dmalloc: Clear file buffers before write: ON Using memory checker: OFF Function Stack Tracing: OFF Strict File Format Checks: OFF Optimization Instrumentation:
Headers are 1.8.15, library is 1.8.18
Your error information shows, then, that you need install the 1.8.15
version.
conda install -c anaconda hdf5=1.8.15
In case someone is having this problem and no amount of uninstalling with pip
or conda
or anything will work: I went through every /lib
path and manually deleted libhdf5*
files, then installed hdf5
with macports
.
Try to uninstall h5py module and install it again. This fixed it for me
pip uninstall h5py
pip install h5py
The only procedure that worked for me on macOS was to create a virtualenv:
virtualenv -p python3 myenv
. myenv/bin/activate
pip3 install h5py==1.10.4 # or whichever version you want
Absolutely nothing else worked!
You may need a brew install hdf5
github issue that did not help:
You can force a custom install of a specific package version:
conda install --force-reinstall anaconda hdf5==1.8.15
or
conda install -c conda-forge hdf5=1.8.15
For some reason uninstalling with conda and then installing hdf5 again with conda is not working. However, if you use pip to uninstall and then conda to install hdf5, it’s working.
So try:
pip uninstall hdf5
conda install hdf5
I tried every method here but only something else entirely worked –
conda install -c conda-forge hdf5=1.10.5 # newer system version
conda install -c conda-forge hdf5=1.8.18 # for this particular problem
This will forcefully install – by either upgrading or downgrading – your hdf5. This in turn will uninstall keras with it and when you reinstall keras that upgrades the hdf5.
My case was trivial and did not need any installations: I had started spyder
and an IPython
console from a conda
environment using cmd
:
activate env
spyder
I then upgraded hdf5
for the env
in cmd
using conda
. However, after running my code in the same IPython
console, I got a similar error message. All it took was to restart spyder
and the corresponding IPython
console. I suppose this reloads the packages consistently.
For people not using conda. As my hdf5 library on Debian was not as new as pip3 supposed, it helped me to install the python3-h5py package using apt as a root, instead of doing it locally via pip3.