I have a dataframe, where I want to get the count of distinct calltypes ((column name).

When I tried using the below spark dataframe transformation code, it gives me data as expected :

fire_df_convrtd_grpby = fire_df_convrtd 
                              .select("CallType") 
                              .where("CallType is not null") 
                              .groupBy("CallType") 
                              .count() 
                              .orderBy("count", ascending=False) 
                              .show()

enter image description here

But when i change he order of the select column, if I have the where method in the begining and select after the count method then the resultant dataset is different

fire_df_convrtd_grpby = fire_df_convrtd 
                            .where("CallType is not null") 
                            .groupBy("CallType") 
                            .count() 
                            .select("CallType") 
                            .orderBy("count", ascending=False) 
                            .show()

enter image description here

why this change of behaviour? as the select coulumn comes in the last does it override the data from count ?