Bugfix: rowgroup filtering by float and double attributes (#28)

adjust · Dec 21, 2020 · 666c49a · 666c49a
1 parent 09ae462
commit 666c49a
Show file tree

Hide file tree

Showing 7 changed files with 53 additions and 17 deletions.
diff --git a/data/example1.parquet b/data/example1.parquet
diff --git a/data/example2.parquet b/data/example2.parquet
diff --git a/data/generate.py b/data/generate.py
@@ -32,7 +32,7 @@
                              date(2018, 1, 5),
                              date(2018, 1, 6)],
                     'six': [False, False, False],
-                    'seven': [0.5, None, 1.0]})
+                    'seven': [1.5, None, 2.0]})
 table2 = pa.Table.from_pandas(df2)
 
 with pq.ParquetWriter('example1.parquet', table1.schema) as writer:

diff --git a/input/parquet_fdw.source b/input/parquet_fdw.source
@@ -42,7 +42,11 @@ SELECT * FROM example1 WHERE one = 2;
 SELECT * FROM example1 WHERE one = 7;
 SELECT * FROM example1 WHERE six = true;
 SELECT * FROM example1 WHERE six = false;
-SELECT * FROM example1 WHERE seven < 0.9;
+SELECT * FROM example1 WHERE seven < 1.5;
+SELECT * FROM example1 WHERE seven <= 1.5;
+SELECT * FROM example1 WHERE seven = 1.5;
+SELECT * FROM example1 WHERE seven > 1;
+SELECT * FROM example1 WHERE seven >= 1;
 SELECT * FROM example1 WHERE seven IS NULL;
 
 -- prepared statements

diff --git a/output/import.source b/output/import.source
@@ -51,10 +51,10 @@ SELECT * FROM example_import ORDER BY one, three;
    2 | {NULL,5,6} | bar   | 2018-01-02 00:00:00 | 2018-01-02 | f   |      
    3 | {7,8,9}    | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
    3 | {21,22}    | zwei  | 2018-01-03 00:00:00 | 2018-01-03 | f   |      
-   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   0.5
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
    5 | {13,14,15} | dos   | 2018-01-05 00:00:00 | 2018-01-05 | f   |      
    5 | {23,24}    | drei  | 2018-01-05 00:00:00 | 2018-01-05 | t   |      
-   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     1
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
    7 | {25,26}    | vier  | 2018-01-07 00:00:00 | 2018-01-07 | f   |      
    9 | {27,28}    | fünf  | 2018-01-09 00:00:00 | 2018-01-09 | t   |      
 (11 rows)

diff --git a/output/parquet_fdw.source b/output/parquet_fdw.source
@@ -24,9 +24,9 @@ SELECT * FROM example1;
    1 | {1,2,3}    | foo   | 2018-01-01 00:00:00 | 2018-01-01 | t   |   0.5
    2 | {NULL,5,6} | bar   | 2018-01-02 00:00:00 | 2018-01-02 | f   |      
    3 | {7,8,9}    | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
-   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   0.5
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
    5 | {13,14,15} | dos   | 2018-01-05 00:00:00 | 2018-01-05 | f   |      
-   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     1
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
 (6 rows)
 
 -- no explicit columns mentions
@@ -93,7 +93,7 @@ SELECT * FROM example1 WHERE one >= 6;
 DEBUG:  parquet_fdw: skip rowgroup 1
  one |    two     | three |        four         |    five    | six | seven 
 -----+------------+-------+---------------------+------------+-----+-------
-   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     1
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
 (1 row)
 
 SELECT * FROM example1 WHERE one = 2;
@@ -122,18 +122,50 @@ SELECT * FROM example1 WHERE six = false;
  one |    two     | three |        four         |    five    | six | seven 
 -----+------------+-------+---------------------+------------+-----+-------
    2 | {NULL,5,6} | bar   | 2018-01-02 00:00:00 | 2018-01-02 | f   |      
-   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   0.5
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
    5 | {13,14,15} | dos   | 2018-01-05 00:00:00 | 2018-01-05 | f   |      
-   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     1
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
 (4 rows)
 
-SELECT * FROM example1 WHERE seven < 0.9;
+SELECT * FROM example1 WHERE seven < 1.5;
+DEBUG:  parquet_fdw: skip rowgroup 2
+ one |   two   | three |        four         |    five    | six | seven 
+-----+---------+-------+---------------------+------------+-----+-------
+   1 | {1,2,3} | foo   | 2018-01-01 00:00:00 | 2018-01-01 | t   |   0.5
+   3 | {7,8,9} | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
+(2 rows)
+
+SELECT * FROM example1 WHERE seven <= 1.5;
  one |    two     | three |        four         |    five    | six | seven 
 -----+------------+-------+---------------------+------------+-----+-------
    1 | {1,2,3}    | foo   | 2018-01-01 00:00:00 | 2018-01-01 | t   |   0.5
-   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   0.5
+   3 | {7,8,9}    | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
+(3 rows)
+
+SELECT * FROM example1 WHERE seven = 1.5;
+DEBUG:  parquet_fdw: skip rowgroup 1
+ one |    two     | three |        four         |    five    | six | seven 
+-----+------------+-------+---------------------+------------+-----+-------
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
+(1 row)
+
+SELECT * FROM example1 WHERE seven > 1;
+DEBUG:  parquet_fdw: skip rowgroup 1
+ one |    two     | three |        four         |    five    | six | seven 
+-----+------------+-------+---------------------+------------+-----+-------
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
 (2 rows)
 
+SELECT * FROM example1 WHERE seven >= 1;
+ one |    two     | three |        four         |    five    | six | seven 
+-----+------------+-------+---------------------+------------+-----+-------
+   3 | {7,8,9}    | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
+(3 rows)
+
 SELECT * FROM example1 WHERE seven IS NULL;
  one |    two     | three |        four         |    five    | six | seven 
 -----+------------+-------+---------------------+------------+-----+-------
@@ -261,9 +293,9 @@ SELECT * FROM example_seq;
    1 | {1,2,3}    | foo   | 2018-01-01 00:00:00 | 2018-01-01 | t   |   0.5
    2 | {NULL,5,6} | bar   | 2018-01-02 00:00:00 | 2018-01-02 | f   |      
    3 | {7,8,9}    | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
-   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   0.5
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
    5 | {13,14,15} | dos   | 2018-01-05 00:00:00 | 2018-01-05 | f   |      
-   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     1
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
    1 | {19,20}    | eins  | 2018-01-01 00:00:00 | 2018-01-01 | t   |      
    3 | {21,22}    | zwei  | 2018-01-03 00:00:00 | 2018-01-03 | f   |      
    5 | {23,24}    | drei  | 2018-01-05 00:00:00 | 2018-01-05 | t   |      
@@ -300,10 +332,10 @@ SELECT * FROM example_sorted ORDER BY one;
    2 | {NULL,5,6} | bar   | 2018-01-02 00:00:00 | 2018-01-02 | f   |      
    3 | {21,22}    | zwei  | 2018-01-03 00:00:00 | 2018-01-03 | f   |      
    3 | {7,8,9}    | baz   | 2018-01-03 00:00:00 | 2018-01-03 | t   |     1
-   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   0.5
+   4 | {10,11,12} | uno   | 2018-01-04 00:00:00 | 2018-01-04 | f   |   1.5
    5 | {23,24}    | drei  | 2018-01-05 00:00:00 | 2018-01-05 | t   |      
    5 | {13,14,15} | dos   | 2018-01-05 00:00:00 | 2018-01-05 | f   |      
-   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     1
+   6 | {16,17,18} | tres  | 2018-01-06 00:00:00 | 2018-01-06 | f   |     2
    7 | {25,26}    | vier  | 2018-01-07 00:00:00 | 2018-01-07 | f   |      
    9 | {27,28}    | fünf  | 2018-01-09 00:00:00 | 2018-01-09 | t   |      
 (11 rows)

diff --git a/parquet_impl.cpp b/parquet_impl.cpp
@@ -2798,9 +2798,9 @@ bytes_to_postgres_type(const char *bytes, arrow::DataType *arrow_type)
         case arrow::Type::INT64:
             return Int64GetDatum(*(int64 *) bytes);
         case arrow::Type::FLOAT:
-            return Int32GetDatum(*(float *) bytes);
+            return Float4GetDatum(*(float *) bytes);
         case arrow::Type::DOUBLE:
-            return Int64GetDatum(*(double *) bytes);
+            return Float8GetDatum(*(double *) bytes);
         case arrow::Type::STRING:
         case arrow::Type::BINARY:
             return CStringGetTextDatum(bytes);