pyspark write.parquet() creates a folder instead of a parquet file

Question

I try to write a pyspark dataframe to a parquet like this

df.write.parquet("temp.parquet", mode="overwrite")

but it creates an empty folder named temp.parquet instead of a parquet file. What might cause this problem?

Answer 1

I downloaded hadoop.dll from here and add it to System32 folder and it solved the problem.