Spark and Spark streaming output difference for the same job

I am doing some POC with Spark and Spark streaming for my project. So all I am doing is reading a file name from Topic. Downloading the file from a "src/main/sresource" and executing the usual "WordCount" frequency application.

 
@KafkaListener(topics = Constants.ABCWordTopic, groupId = Constants.ABC_WORD_COMSUMER_GROUP_ID) 
public void processTask(@Payload String fileResourcePath) {
        log.info("ABC Receiving task from WordProducer filepath {} at time {}", fileResourcePath,
                LocalDateTime.now());
        // Spark job
        /*
         * JavaRDD wordRDD =
         * sparkContext.parallelize(Arrays.asList(extractFile(fileResourcePath).split(" ")));
         * log.info("ABC Map Contents : {}", wordRDD.countByValue().toString());
         * wordRDD.coalesce(1,
         * true).saveAsTextFile("ResultSparklog_"+ System.currentTimeMillis());
         */
        // Spark Streaming job
        JavaPairDStream wordPairStream = streamingContext
                .textFileStream(extractFile(fileResourcePath))
                .flatMap(line -> Arrays.asList(SPACE.split(line)).iterator())
                .mapToPair(s -> new Tuple2(s, 1)).reduceByKey((i1, i2) -> i1 + i2);
        wordPairStream.foreachRDD(wordRDD -> {
        //  javaFunctions(wordTempRDD).writerBuilder("vocabulary", "words", mapToRow(String.class))
        //                  .saveToCassandra();
            log.info("ABC Map Contents : {}", wordRDD.keys().countByValue().toString());
            wordRDD.coalesce(1, true)
                    .saveAsTextFile("SparkStreamResultlog_" + System.currentTimeMillis());
        });
        streamingContext.start();
        try {
            streamingContext.awaitTerminationOrTimeout(-1);
        } catch (InterruptedException e) {
            log.error("Terminated streaming context {}", e);
        }
    }

• In the above code I am listening from a Kafka Topic("ABCtopic") and
  processing it. The "Spark job" commented code works perfectly fine.
  It counts the word and gives the results as expected, however "spark
  streaming job" code does not behave as expected and it outputs null.
  
• The line log.info("ABC Map Contents : {}", wordRDD.keys().countByValue().toString()); gives '{}' as output.
  Writing to a file is empty. Being new to Spark streaming from what
  little in know "Spark streaming" is an additional library for
  continuously processing data at real time from any source like a
  file, topic etc.
• What is missing in the above code for the spark streaming to output
  'null' in the highlighted log line and also in the output data file
  which is being written to disk whereas the Spark job does the same
  job perfectly fine.