load save in spark

Question

1) load command will load only parquet 

val a = spark.read.load("employee.parquet")--works fine
val a=spark.read.load("employee.txt")--error

2)val a =spark.read.format("csv").load("employee.txt")--works fine

3)what is the difference between 1 and 2 point loading text as data 

4)except parquet file remaining format are not accepting the below stmt 

val a = spark.read.load("text/csv/json") 

5) what is the difference of this two stmts 

val a =spark.read.json()
val a =spark.read.format("json").load("a.json")

Answer 1

The reason why you are able to load employee.parquet and not employee.txt using load in spark.read.load by default it assumes that data source is in parquet format so it is able to load it but we can use format function which can be used to specify the different format and use the load function to load the data 

spark.read.json()

Spark SQL can automatically infer the schema of a JSON dataset and load it as a Dataset[Row]. This conversion can be done by spark.read.json()

Both spark.read.json() and spark.read.format("json").load("a.json") are same as you can see in the below screenshot,