How does Spark handle Timestamp types during Pandas dataframe conversion?

I have a pandas dataframe with timestamp columns of type pandas.tslib.Timestamp. I looked through the pyspark source code from 'createDataFrame'(link to source) and it seems that they convert the data to a numpy record array to a list:

data = [r.tolist() for r in data.to_records(index=False)]

However, timestamp types get converted in this process to a list of longs:

> df = pd.DataFrame(pd.date_range(start=datetime.datetime.now(),periods=5,freq='s'))
> df
0 2017-07-25 11:53:29.353923
1 2017-07-25 11:53:30.353923
2 2017-07-25 11:53:31.353923
3 2017-07-25 11:53:32.353923
4 2017-07-25 11:53:33.353923
> df.to_records(index=False).tolist()
[(1500983799614193000L,), (1500983800614193000L,), (1500983801614193000L,), (1500983802614193000L,), (1500983803614193000L,)]

Now if I pass such a list to a RDD, do some operations(not touching the timestamp column) and then call

> spark.createDataFrame(rdd,schema) // with schema mentioning that column as TimestampType
TypeError: TimestampType can not accept object 1465197332112000000L in type <type 'long'>
    at org.apache.spark.api.python.PythonRunner$$anon$1.read(PythonRDD.scala:193)
    at org.apache.spark.api.python.PythonRunner$$anon$1.<init>(PythonRDD.scala:234)
    at org.apache.spark.api.python.PythonRunner.compute(PythonRDD.scala:152)
    at org.apache.spark.api.python.PythonRDD.compute(PythonRDD.scala:63)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)
    at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
    at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:323)
    at org.apache.spark.rdd.RDD.iterator(RDD.scala:287)
    at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:87)
    at org.apache.spark.scheduler.Task.run(Task.scala:99)
    at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:322)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)

What should I do(before converting the list to a RDD) to preserve the datetime type.

Edit 1

A few methods I'm aware of that will involve post dataframe creation processing are:

1. Add timezone information to datetime object in pandas. However this seems unnecessary and can lead to errors depending on your own timezone.

2. Convert the long to a timestamp using the datetime library.

Assuming tstampl is the input: tstamp = datetime(1970, 1, 1) + timedelta(microseconds=tstampl/1000)

3. Convert the datetime to string on Pandas dataframe side, then cast to datetime on Spark dataframe side.

As explained in Suresh's answer below

However I'm looking for a simpler way which would take care of all processing before dataframe creation itself.