[MADLIB-1186] Correlation query fails with international chars in column name

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: v1.13
• Component/s: Module: Inferential Statistics
• Assignee: ssoni
• Reporter: Nandish Jayaram
• Created: 11/Dec/17 22:03
• Updated: 13/Dec/17 19:58
• Resolved: 13/Dec/17 19:58

Description

We tried running the correlation module in MADlib, on a table that had
international characters for schema, and column name of one of the
independent variables. It resulted in the following error:

select
      madlib.correlation(
          '"nåmespace"."dt_gOlf"',
          '"__madlib_temp_26728701_1513000192_9340672___πøA"',
          'id, temperature, "Humidity"');
psql:/tmp/build/60928da1/madlib_testsuite/tests/Correlation/sql_CorrelationInternationalCharOutputTestCase/test_correlation_international_char.sql:12: ERROR:  spiexceptions.SyntaxError: syntax error at or near ""Humidity""
LINE 10: ...rature, avg_temperature),coalesce("Humidity", avg_"Humidity"...
                                                              ^
QUERY:

            CREATE TEMP TABLE __madlib_temp_43456465_1513000195_75531635__ AS
            SELECT
                count(*) AS tot_cnt,
                mean,
                madlib.correlation_agg(x, mean) as cor_mat
            FROM
            (
                SELECT ARRAY[ coalesce(id, avg_id),coalesce(temperature, avg_temperature),coalesce("Humidity", avg_"Humidity") ] AS x,
                        ARRAY [ avg_id,avg_temperature,avg_"Humidity" ] AS mean
                FROM "nåmespace"."dt_gOlf",
                (
                    SELECT avg(id) AS avg_id,avg(temperature) AS avg_temperature,avg("Humidity") AS avg_"Humidity"
                    FROM "nåmespace"."dt_gOlf"
                )sub1
            ) sub2
            GROUP BY mean

CONTEXT:  Traceback (most recent call last):
  PL/Python function "correlation", line 23, in <module>
    return correlation.correlation(**globals())
  PL/Python function "correlation", line 71, in correlation
  PL/Python function "correlation", line 205, in _populate_output_table
PL/Python function "correlation"
      select * from "__madlib_temp_26728701_1513000192_9340672___πøA" order by column_position;

Issue Links

• links to: GitHub Pull Request #214