- last edited on
I run Spark Streaming to fetch all files from directory in batches.
If the a file with the same name were copied to the directory - that Sparks look for - and its modified date is within the window of the batch streaming, Spark ignores it.
Any other new files with different name in the directory, and its modified date within the window of the batch streaming, Spark fetch normally
This contradicts with the documentation were spark state they look for modification date only.
Spark Streamin - textFileStream
Anyone has explaination?