Spark incremental loading overwrite old record

Question

I have a requirement to do the incremental loading to a table by using Spark (PySpark)

Here\'s the example:

Day 1

id | value
-----------
1  |         


        

Answer 1

dataframe appending is done by union function in pyspark. I'll demo with an example and create 2 dataframes as you mentioned in the question.

from pyspark.sql.types import Row
df1 = sqlContext.createDataFrame([Row(id=1,value="abc"),Row(id=2,value="def")])

df1.show()
+---+-----+
| id|value|
+---+-----+
|  1|  abc|
|  2|  def|
+---+-----+

df2 = sqlContext.createDataFrame([Row(id=2,value="cde"),Row(id=3,value="xyz")])
df2.show()
+---+-----+
| id|value|
+---+-----+
|  2|  cde|
|  3|  xyz|
+---+-----+

Lets do a union between the two dataframes and you will get the desired result.

df2.union(df1).dropDuplicates(["id"]).show()
+---+-----+
| id|value|
+---+-----+
|  1|  abc|
|  3|  xyz|
|  2|  cde|
+---+-----+

You can sort the output using asc from pyspark.sql.functions

from pyspark.sql.functions import asc


df2.union(df1).dropDuplicates(["id"]).sort(asc("id")).show()
+---+-----+
| id|value|
+---+-----+
|  1|  abc|
|  2|  cde|
|  3|  xyz|
+---+-----+