[SPARK-19386][SparkR][Followup] fix error in vignettes

[actuaryzhang]

What changes were proposed in this pull request?

Current version has error in vignettes:

model <- spark.bisectingKmeans(df, Sepal_Length ~ Sepal_Width, k = 4)
summary(kmeansModel)

kmeansModel does not exist...

@felixcheung @wangmiao1981

[actuaryzhang]

Errors can be seen in #16794.

[wangmiao1981]

Good catch! Thanks for fixing it!

@felixcheung This is a hot-fix to build break. I don't know why Jenkins didn't catch it in the original PR.

[wangmiao1981]

Jenkins, test this please.

[wangmiao1981]

LGTM. Let us wait for Jenkins to build it.

[SparkQA]

Test build #72338 has finished for PR 16799 at commit 9552f1f.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[gatorsmile]

Since this PR is a HOT fix, let me merge it now. Thanks!

[gatorsmile]

Thanks! Merging to master.

[SparkQA]

Test build #72342 has finished for PR 16799 at commit 9552f1f.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[felixcheung]

thanks @gatorsmile, @actuaryzhang @wangmiao1981
we did see it on the other PR but I didn't put the two together. And sorry for missing the fact that Jenkins was never run on the PR that broken the build - comment threads were too long and AppVeyor green checkmark was there (need to follow up) - will be more careful the next time.