[SPARK-29034][SQL] PostgreSQL dialect conformance for string constants with C-style escapes

[yaooqinn]

What changes were proposed in this pull request?

On one hand, we now use spark.sql.parser.escapedStringLiterals to control whether to escape string literal or not.
On the other hand, we use spark.sql.dialect to choose spark or PostgreSQL dialect. When we use the PostgreSQL dialect, we should obey the C-style escape behavior of PostgreSQL.

Supported

An escape string constant is specified by writing the letter E (upper or lower case) just before the opening single quote, e.g., E'foo'.

Not supported

When continuing an escape string constant across lines, write E only before the first opening quote.

Because PostgreSQL follows the SQL standard that is

Two string constants that are only separated by whitespace with at least one newline are concatenated and effectively treated as if the string had been written as one constant.

, which is hard to follow in Spark's Parser.

Why are the changes needed?

PostgreSQL dialect conformance

Does this PR introduce any user-facing change?

yes, when we use the PostgreSQL dialect, we use 'E' to define an escape string constant

How was this patch tested?

add ut.

[yaooqinn]

retest this please

[SparkQA]

Test build #114629 has finished for PR 26715 at commit 7ae10e9.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[SparkQA]

Test build #114634 has finished for PR 26715 at commit db93221.

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

[yaooqinn]

closed because pg dialect removed