Inserting documents in MongoDB with Spark Connector (Dataframe vs Spark Structured Streaming)

MongoDBsparkConnector.scala

// Dataframe (supported) - read 1 file, no streaming
// Step 1, create the Dataframe source
val fileDF = spark
  .read // No streaming
  .csv("file/file1.csv")
  .selectExpr("CAST(key as String)", // more code with other casting...
              )
  // Out [1]: fileDF: org.apache.spark.sql.package.DataFrame = [key: string, country: string ... 6 more fields]

// Step 2, insert Dataframe into MongoDB
MongoSpark.save(fileDF.write.option("collection", "trends").mode("append"))
  // inserts with no problem all the documents contained at the Dataframe with the correct structure

//////////////////////////////////////////////////////////////////////////////////
//////////////////////////////////////////////////////////////////////////////////

// Spark Structured Streaming (not supported) - read multiple files, streaming
// Step 1, create the Dataframe source
val filesDS = spark
  .readStream // Streaming
  .csv("trendFiles/*.csv")
  .selectExpr("CAST(key as String)", // more code with other casting...
              )
  // Out [2]: filesDS: org.apache.spark.sql.package.DataFrame = [key: string, country: string ... 6 more fields]

// Step 2, insert Dataframe into MongoDB
// I cannot see MongoSpark.save method for Stuctured Streaming.

I opened a ticket in MongoDB's Jira some years ago:
https://jira.mongodb.org/browse/SPARK-134

You can see the details in the comments. I did not test it but it seems the solution is to use MongoConnector directly (Ross Lawley updated master and Christiaan Ras used it with no problem). Sorry if I could not provide more useful help.

Many thanks, Amy example for structured streaming using python(pyspark)?

No, sorry, I did not code structured streaming using Python.