Use pre-existing files in AbstractTestParquetPageSkipping

Currently optimized parquet writer does not support writing
page indexes. Using files prepared from the legacy writer for
these tests allows having test coverage for page indexes after
removal of the legacy writer.

plugin/trino-hive/src/test/java/io/trino/plugin/hive/AbstractTestParquetPageSkipping.java

@@ -13,6 +13,7 @@
  */
 package io.trino.plugin.hive;
 
+import com.google.common.io.Resources;
 import io.trino.Session;
 import io.trino.execution.QueryStats;
 import io.trino.operator.OperatorStats;
@@ -27,6 +28,8 @@
 import org.testng.annotations.DataProvider;
 import org.testng.annotations.Test;
 
+import java.io.File;
+import java.net.URISyntaxException;
 import java.util.Map;
 
 import static com.google.common.collect.MoreCollectors.onlyElement;
@@ -77,10 +80,29 @@ private void buildSortedTables(String tableName, String sortByColumnName, String
     }
 
     @Test
-    public void testAndPredicates()
+    public void testRowGroupPruningFromPageIndexes()
+            throws Exception
     {
-        String tableName = "test_and_predicate_" + randomNameSuffix();
-        buildSortedTables(tableName, "totalprice", "double");
+        String tableName = "test_row_group_pruning_" + randomNameSuffix();
+        File parquetFile = new File(Resources.getResource("parquet_page_skipping/orders_sorted_by_totalprice").toURI());
+        assertUpdate(
+                """
+                        CREATE TABLE %s (
+                           orderkey bigint,
+                           custkey bigint,
+                           orderstatus varchar(1),
+                           totalprice double,
+                           orderdate date,
+                           orderpriority varchar(15),
+                           clerk varchar(15),
+                           shippriority integer,
+                           comment varchar(79),
+                           rvalues double array)
+                        WITH (
+                           format = 'PARQUET',
+                           external_location = '%s')
+                        """.formatted(tableName, parquetFile.getAbsolutePath()));
+
         int rowCount = assertColumnIndexResults("SELECT * FROM " + tableName + " WHERE totalprice BETWEEN 100000 AND 131280 AND clerk = 'Clerk#000000624'");
         assertThat(rowCount).isGreaterThan(0);
 
@@ -92,19 +114,14 @@ public void testAndPredicates()
 
     @Test
     public void testPageSkippingWithNonSequentialOffsets()
+            throws URISyntaxException
     {
         String tableName = "test_random_" + randomNameSuffix();
-        int updateCount = 8192;
-        assertUpdate(
-                "CREATE TABLE " + tableName + " (col) WITH (format = 'PARQUET') AS " +
-                        "SELECT * FROM unnest(transform(repeat(1, 8192), x -> rand()))",
-                updateCount);
-        for (int i = 0; i < 8; i++) {
-            assertUpdate(
-                    "INSERT INTO " + tableName + " SELECT rand() FROM " + tableName,
-                    updateCount);
-            updateCount += updateCount;
-        }
+        File parquetFile = new File(Resources.getResource("parquet_page_skipping/random").toURI());
+        assertUpdate(format(
+                "CREATE TABLE %s (col double) WITH (format = 'PARQUET', external_location = '%s')",
+                tableName,
+                parquetFile.getAbsolutePath()));
         // These queries select a subset of pages which are stored at non-sequential offsets
         // This reproduces the issue identified in https://github.com/trinodb/trino/issues/9097
         for (double i = 0; i < 1; i += 0.1) {
@@ -115,12 +132,17 @@ public void testPageSkippingWithNonSequentialOffsets()
 
     @Test
     public void testFilteringOnColumnNameWithDot()
+            throws URISyntaxException
     {
         String nameInSql = "\"a.dot\"";
         String tableName = "test_column_name_with_dot_" + randomNameSuffix();
 
-        assertUpdate("CREATE TABLE " + tableName + "(key varchar(50), " + nameInSql + " varchar(50)) WITH (format = 'PARQUET')");
-        assertUpdate("INSERT INTO " + tableName + " VALUES ('null value', NULL), ('sample value', 'abc'), ('other value', 'xyz')", 3);
+        File parquetFile = new File(Resources.getResource("parquet_page_skipping/column_name_with_dot").toURI());
+        assertUpdate(format(
+                "CREATE TABLE %s (key varchar(50), %s varchar(50)) WITH (format = 'PARQUET', external_location = '%s')",
+                tableName,
+                nameInSql,
+                parquetFile.getAbsolutePath()));
 
         assertQuery("SELECT key FROM " + tableName + " WHERE " + nameInSql + " IS NULL", "VALUES ('null value')");
         assertQuery("SELECT key FROM " + tableName + " WHERE " + nameInSql + " = 'abc'", "VALUES ('sample value')");
@@ -157,17 +179,15 @@ public void testPageSkipping(String sortByColumn, String sortByColumnType, Objec
 
     @Test
     public void testFilteringWithColumnIndex()
+            throws URISyntaxException
     {
         String tableName = "test_page_filtering_" + randomNameSuffix();
-        String catalog = getSession().getCatalog().orElseThrow();
-        assertUpdate(
-                Session.builder(getSession())
-                        .setCatalogSessionProperty(catalog, "parquet_writer_page_size", "32kB")
-                        .build(),
-                "CREATE TABLE " + tableName + " " +
-                        "WITH (format = 'PARQUET', bucket_count = 1, bucketed_by = ARRAY['suppkey'], sorted_by = ARRAY['suppkey']) AS " +
-                        "SELECT suppkey, extendedprice, shipmode, comment FROM tpch.tiny.lineitem",
-                60175);
+        File parquetFile = new File(Resources.getResource("parquet_page_skipping/lineitem_sorted_by_suppkey").toURI());
+        assertUpdate(format(
+                "CREATE TABLE %s (suppkey bigint, extendedprice decimal(12, 2), shipmode varchar(10), comment varchar(44)) " +
+                        "WITH (format = 'PARQUET', external_location = '%s')",
+                tableName,
+                parquetFile.getAbsolutePath()));
 
         verifyFilteringWithColumnIndex("SELECT * FROM " + tableName + " WHERE suppkey = 10");
         verifyFilteringWithColumnIndex("SELECT * FROM " + tableName + " WHERE suppkey BETWEEN 25 AND 35");