Consider the following Spark Code to construction a Dataframe.

a $=[(\text{'}$Chris$\text{'},$ 'Budweiser', $15),(\text{'}$Chris$\text{'},$ 'Becks', $5),(\text{'}$Chris$\text{'},$ 'Heineken', $2),(\text{'}$Bob$\text{'},$ 'Becks', $15),(\text{'}$Bob$\text{'},$ 'Budweiser', $10),(\text{'}$Bob$\text{'},$

'Heineken', $2),(\text{'}$Alice$\text{'},$ 'Heineken', $8)]$

rdd $=$ sc$.$ parallelize $(a)$

df $=$ sqIContext.createDataFrame$($rdd$,[\text{'}$drinker$\text{'},$ 'beer', 'score'$])$

sqIContext.registerDataFrameAsTable$($df$,$ "drinkers"$)$

How can we get the total score of each beer brand?

We want to have the following answer from the above example: $($Your print out may be different $-$ values are important$)$

beer$=$'Becks', total score $=20$

beer$=$'Budweiser', total score$=25$

beer$=$'Heineken', total score$=12$

$($Multiple Choice with Negative Scores for wrong Answers$)$

A$.$ df$.$drop$(\text{'}$drinker$\text{'}).$groupByKey$(\text{'}$beer$\text{'}).$reduceByKey$($add$).$collect$()$

B$.$ df$.$drop$(\text{'}$drinker$\text{'}).$groupBy$(\text{'}$beer$\text{'}).$agg$(\{\text{'}$score$\text{'}$: 'sum'$\}).$collect$()$

C$.$ df$.$drop$(\text{'}$drinker$\text{'}).$groupByKey$(\text{'}$beer$\text{'}).$map$($lambda a$,$ b: a$+$b$).$top$()$

D$.$ df$.$filter$(\text{'}$drinker$\text{'}).$groupBy$(\text{'}$beer$\text{'}).$agg$(\{\text{'}$score$\text{'}$: 'sum'$\}).$collect$()$

E$.$ df$.$filter$(\text{'}$drinker$\text{'}).$groupByKey$(\text{'}$beer$\text{'}).$reduceByKey$($add$).$collect$()$

F$.$ sqIContext.sql$("$SELECT beer, sum$($score$)$ from drinkers GROUP BY drinker"$).$collect$()$

G$.$ df$.$filter$(\text{'}$drinker$\text{'}).$groupByKey$(\text{'}$beer$\text{'}).$agg$(\{\text{'}$score$\text{'}$: 'sum'$\}).$collect$()$

H$.$ df$.$drop$(\text{'}$drinker$\text{'}).$groupByKey$(\text{'}$beer$\text{'}).$map$($lambda a$,$ b: a$+$b$).$collect$()$

I. sqIContext.sqI$("$SELECT beer, sum$($score$)$ from drinkers GROUP BY beer"$).$collect$()$