为什么 xgboost 工作/不工作取决于其他软件包的导入顺序？

Question

当尝试使用以下命令导入 xgboost 时：

import pandas as pd,  numpy as np, psycopg2, xgboost as xgb, os, shutil, pickle, random

我收到以下错误：

  File "/opt/virtual_environments/ml_venv/lib/python3.6/site-packages/xgboost/__init__.py", line 9, in <module>
    from .core import DMatrix, DeviceQuantileDMatrix, Booster, DataIter
  File "/opt/virtual_environments/ml_venv/lib/python3.6/site-packages/xgboost/core.py", line 203, in <module>
    _LIB = _load_lib()
  File "/opt/virtual_environments/ml_venv/lib/python3.6/site-packages/xgboost/core.py", line 194, in _load_lib
    """)
xgboost.core.XGBoostError: 
XGBoost Library (libxgboost.so) could not be loaded.
Likely causes:
  * OpenMP runtime is not installed
    - vcomp140.dll or libgomp-1.dll for Windows
    - libomp.dylib for Mac OSX
    - libgomp.so for Linux and other UNIX-like OSes
    Mac OSX users: Run `brew install libomp` to install OpenMP runtime.

  * You are running 32-bit Python on a 64-bit OS

Error message(s): ['dlopen: cannot load any more object with static TLS']

在 psycopg2 前面切换 xgboost 的导入，如下所示：

import pandas as pd,  numpy as np, xgboost as xgb, psycopg2, os, shutil, pickle, random

代码导入，找到共享库并运行良好。

我知道这可能取决于服务器/环境的设置，但是首先导入 xgboost 时它的工作是否有某种原因？ psycopg2 是否可能更改系统路径或环境变量？

干杯