[DRILL-5930] Table function escape handling does not handle lone backslash

• Type: Bug
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Affects Version/s: 1.11.0
• Assignee: Paul Rogers
• Reporter: Paul Rogers
• Created: 03/Nov/17 23:18
• Updated: 21/Jan/19 13:43

Description

Consider the following query from the test framework (Functional/table_function/positive/drill-3149_10.q):

select * from table(`table_function/colons.txt`(type=>'text',lineDelimiter=>'\\'))

Notice the double-backslash. Drill translates this into a single backslash. But, what happens if the user provides just a single backslash? Drill translates that into the empty string. Drill passes the empty string to the text reader, which fails with the error described in DRILL-5929.

Better would be to either:

• Leave lone back-slashes unchanged, or
• Fail a query with an lone backslash.

The fix for this bug prefers the first (because it is easiest.) However, this fix does not handle a triple backslash, which will be translated to a single backslash:

• \ \ --> \
• \ --> (nothing)

The code in question:

class FormatPluginOp ...
  FormatPluginConfig createConfigForTable(TableInstance t) {
...
          param = StringEscapeUtils.unescapeJava(param);

Issue Links

• Is contained by: DRILL-6986 Table function improvements / issues (UMBRELLA JIRA)