enable udf in nexmark q14

Signed-off-by: Runji Wang <[email protected]>
risingwavelabs · Feb 4, 2024 · 7b640e2 · 7b640e2
1 parent f5386ca
commit 7b640e2
Show file tree

Hide file tree

Showing 13 changed files with 41 additions and 31 deletions.
diff --git a/ci/scripts/sql/nexmark/q14.drop.sql b/ci/scripts/sql/nexmark/q14.drop.sql
@@ -1,3 +1,4 @@
 -- noinspection SqlNoDataSourceInspectionForFile
 -- noinspection SqlResolveForFile
 DROP SINK nexmark_q14;
+DROP FUNCTION count_char;
diff --git a/ci/scripts/sql/nexmark/q14.sql b/ci/scripts/sql/nexmark/q14.sql
@@ -1,5 +1,12 @@
 -- noinspection SqlNoDataSourceInspectionForFile
 -- noinspection SqlResolveForFile
+
+CREATE FUNCTION count_char(s varchar, c varchar) RETURNS int LANGUAGE rust AS $$
+  fn count_char(s: &str, c: &str) -> i32 {
+      s.matches(c).count() as i32
+  }
+$$;
+
 CREATE SINK nexmark_q14 AS
 SELECT auction,
        bidder,
@@ -15,11 +22,8 @@ SELECT auction,
                THEN 'nightTime'
            ELSE 'otherTime'
            END       AS bidTimeType,
-       date_time
-       -- extra
-       -- TODO: count_char is an UDF, add it back when we support similar functionality.
-       -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
-       -- count_char(extra, 'c') AS c_counts
+       date_time,
+       count_char(extra, 'c') AS c_counts
 FROM bid
 WHERE 0.908 * price > 1000000
   AND 0.908 * price < 50000000

diff --git a/e2e_test/streaming/nexmark/sinks/q14.slt.part b/e2e_test/streaming/nexmark/sinks/q14.slt.part
@@ -1,3 +1,10 @@
+statement ok
+CREATE FUNCTION count_char(s varchar, c varchar) RETURNS int LANGUAGE rust AS $$
+  fn count_char(s: &str, c: &str) -> i32 {
+      s.matches(c).count() as i32
+  }
+$$;
+
 statement ok
 CREATE SINK nexmark_q14 AS
 SELECT
@@ -15,11 +22,8 @@ SELECT
     THEN 'nightTime'
     ELSE 'otherTime'
   END AS bidTimeType,
-  date_time
-  -- extra
-  -- TODO: count_char is an UDF, add it back when we support similar functionality.
-  -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
-  -- count_char(extra, 'c') AS c_counts
+  date_time,
+  count_char(extra, 'c') AS c_counts
 FROM bid
 WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000
 WITH ( connector = 'blackhole', type = 'append-only', force_append_only = 'true');
diff --git a/e2e_test/streaming/nexmark/views/q14.slt.part b/e2e_test/streaming/nexmark/views/q14.slt.part
@@ -1,3 +1,10 @@
+statement ok
+CREATE FUNCTION count_char(s varchar, c varchar) RETURNS int LANGUAGE rust AS $$
+  fn count_char(s: &str, c: &str) -> i32 {
+      s.matches(c).count() as i32
+  }
+$$;
+
 statement ok
 CREATE MATERIALIZED VIEW nexmark_q14 AS
 SELECT
@@ -15,10 +22,7 @@ SELECT
     THEN 'nightTime'
     ELSE 'otherTime'
   END AS bidTimeType,
-  date_time
-  -- extra
-  -- TODO: count_char is an UDF, add it back when we support similar functionality.
-  -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
-  -- count_char(extra, 'c') AS c_counts
+  date_time,
+  count_char(extra, 'c') AS c_counts
 FROM bid
 WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;
diff --git a/e2e_test/udf/wasm/src/lib.rs b/e2e_test/udf/wasm/src/lib.rs
@@ -1,6 +1,11 @@
 use arrow_udf::function;
 use rust_decimal::Decimal;
 
+#[function("count_char(varchar, varchar) -> int")]
+fn count_char(s: &str, c: &str) -> i32 {
+    s.matches(c).count() as i32
+}
+
 #[function("int_42() -> int")]
 fn int_42() -> i32 {
     42

diff --git a/src/frontend/planner_test/tests/testdata/input/nexmark.yaml b/src/frontend/planner_test/tests/testdata/input/nexmark.yaml
@@ -336,8 +336,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/input/nexmark_source.yaml b/src/frontend/planner_test/tests/testdata/input/nexmark_source.yaml
@@ -338,8 +338,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/input/nexmark_temporal_filter.yaml b/src/frontend/planner_test/tests/testdata/input/nexmark_temporal_filter.yaml
@@ -321,8 +321,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/input/nexmark_watermark.yaml b/src/frontend/planner_test/tests/testdata/input/nexmark_watermark.yaml
@@ -305,8 +305,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/output/nexmark.yaml b/src/frontend/planner_test/tests/testdata/output/nexmark.yaml
@@ -1185,8 +1185,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/output/nexmark_source.yaml b/src/frontend/planner_test/tests/testdata/output/nexmark_source.yaml
@@ -1003,8 +1003,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/output/nexmark_temporal_filter.yaml b/src/frontend/planner_test/tests/testdata/output/nexmark_temporal_filter.yaml
@@ -957,8 +957,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;

diff --git a/src/frontend/planner_test/tests/testdata/output/nexmark_watermark.yaml b/src/frontend/planner_test/tests/testdata/output/nexmark_watermark.yaml
@@ -1208,8 +1208,7 @@
       END AS bidTimeType,
       date_time,
       extra
-      -- TODO: count_char is an UDF, add it back when we support similar functionality.
-      -- https://github.com/nexmark/nexmark/blob/master/nexmark-flink/src/main/java/com/github/nexmark/flink/udf/CountChar.java
+      -- ignore UDF in planner test
       -- count_char(extra, 'c') AS c_counts
     FROM bid
     WHERE 0.908 * price > 1000000 AND 0.908 * price < 50000000;