[SPARK-25517][SQL] fixes for auto inferSchema for date columns as date instead of string

[ghost]

This working patch is for [SPARK-25517][SQL] which is about auto inferring date type from the csv dataset while loading the csv dataset into a dataframe, as per the current behavior it reads date column as string type which is wrong it should read as date. however auto-inferring schema for timestamp column is perfectly working fine

What changes were proposed in this pull request?

Modified a single file CSVInferSchema.scala and made changes in the corresponding Suite i.e. CSVInferSchemaSuite and run this suite and it's working perfectly now,

How was this patch tested?

Performed unit testing using the suite CSVInferSchemaSuite.scala

[ghost]

Hi @dongjoon-hyun @HyukjinKwon

Regret for being inactive for sometime, this is going to be my first contribution to the spark realm, I am honestly totally new to github and self-experimenting to learn how to operate through github. hence created a separate github account to work on this jira [SPARK-25517] and raised this PR.

over the last weekend, I verified that this issue still exists in the latest spark 2.4.0 hence worked on it and fixed the issue. Updated the test suite CSVInferSchemaSuite.scala and tested through it, working fine now as expected,

so I'd like to reopen this jira and mark this as closed,

please review the changes and suggest, in case of any further inputs or queries you are most welcome, : )

Regards,
Manoranjan

[MaxGekk]

Similar changes were committed #23202 and reverted by @HyukjinKwon already. @manoranjanmix Welcome here and good luck.

[HyukjinKwon]

Yup, please make sure of addressing the concerns made there.

[ghost]

Ohh i see that now Max, thanks for pointing, I had caught this issue way back in Sep-18 but was unfortunate to continue, however I will proceed further from here and try to bring more improvements
Thanks,

[AmplabJenkins]

Can one of the admins verify this patch?

[HyukjinKwon]

Closing this due to author's inactivity.