Conda Environment Variable discrepancy

Product Help Packages & Environments
1

Hello everyone. I have a conda environment I am setting up with kubernetes and pyspark. The initial issue is derived from the below, when trying to run pyspark.

(kpyEnv) C:\WINDOWS\system32>pyspark
The system cannot find the path specified.

After much digging in places like here , I figured it was a dependent item or environment variable. I verified the downloads, and that my environment variables for Hadoop (HADOOP_HOME), java (JAVA_HOME), and spark (SPARK_HOME) were all correct in the system. But when I echo JAVA_HOME inside of my conda environment, it gives me a path to \myEnvironmentName\Library
I am unsure why. I tried manually setting the path inside of the environment but despite the output in both a standard command prompt as well as conda-meta\state being the same as my Windows environment variable path, the “Echo %JAVA_HOME%” command still results in leading to the environment library. I also get a warning when starting about “overwriting environment variables set in the machine overwriting variable {‘JAVA_HOME’}” on conda activate, and yet it’s still wrong.

After hours of troubleshooting I created an account here (after browsing some topics, yes. I found a similar one possibly that went unanswered but that was it.)

2

Edit - I finally solved it. There was a second .bat file that was overriding my environment variable AFTER it was being set every time the environment was created. Unfortunately this didn’t solve the pyspark issue, but it at least resolved this ticket.

3

hello, I also met the same problems. When I source conda, I got that:
“WARNING: overwriting environment variables set in the machine
overwriting variable {‘PS1’}” .
How did you solve it? It will be kind if you can help me. Thank you!