test: exercise everyday SQL expressions in the sample and on PostgreSQL

examples/maven-postgresql/sql/migrations/V4__add_author_updated_at.sql

@@ -0,0 +1 @@
+ALTER TABLE authors ADD COLUMN updated_at TIMESTAMP;

examples/maven-postgresql/sql/queries.sql

@@ -64,3 +64,33 @@ SELECT *
 FROM books
 WHERE format = $1
 ORDER BY id;
+
+-- name: FilterAuthors :many
+SELECT *
+FROM authors
+WHERE (:name::text IS NULL OR name = :name)
+ORDER BY id;
+
+-- name: RenameAuthor :exec
+UPDATE authors
+SET name = :name, updated_at = now()
+WHERE id = :id;
+
+-- name: CountBooksByAuthor :many
+SELECT author_id, COUNT(*) AS books
+FROM books
+GROUP BY author_id
+ORDER BY author_id;
+
+-- name: UpsertAuthor :one
+INSERT INTO authors (id, name, bio)
+VALUES (:id, :name, :bio)
+ON CONFLICT (id) DO UPDATE
+    SET name = EXCLUDED.name, bio = EXCLUDED.bio, updated_at = now()
+RETURNING *;
+
+-- name: ListAuthorsByIds :many
+SELECT *
+FROM authors
+WHERE id = ANY(:ids)
+ORDER BY id;

examples/maven-postgresql/src/main/java/com/example/app/App.java

@@ -21,22 +21,26 @@
  * code: create, read, optional read, list, update, a missing-row read, a
  * committed transaction, a rolled back transaction, a case-insensitive search,
  * a count, a page, a book insert, a left-joined projection, a cataloged book
- * insert, a read by enum value, and delete.
+ * insert, a read by enum value, an optional filter by name, a read by id list,
+ * a grouped book count, a rename, an upsert, and delete.
  *
- * <p>All thirteen named queries are methods of one generated
+ * <p>All eighteen named queries are methods of one generated
  * {@link AuthorRepository}. The repository is constructed once per execution
  * context: once from the {@code DataSource}-backed executor, and once more per
  * transaction from a caller-owned connection.
  *
- * <p>Create, read, optional read, list, search, and page each return one
- * complete {@code authors} row, so all six share the single top-level
- * {@link AuthorsRow} record of the generated package, and the two queries that
- * return a complete {@code books} row share the top-level {@link BooksRow}
- * record. A query with its own result shape generates its own nested record:
- * {@link AuthorRepository.CountAuthorsResult} carries the non-null
- * {@code Long} count, and {@link AuthorRepository.ListAuthorBooksResult}
- * carries an author name beside the title of the left-joined {@code books} row,
- * which is {@code null} for an author without a book.
+ * <p>Create, read, optional read, list, search, page, filter, upsert, and the
+ * read by id list each return one complete {@code authors} row, so all nine
+ * share the single top-level {@link AuthorsRow} record of the generated
+ * package, and the two queries that return a complete {@code books} row share
+ * the top-level {@link BooksRow} record. A query with its own result shape
+ * generates its own nested record: {@link AuthorRepository.CountAuthorsResult}
+ * carries the non-null {@code Long} count,
+ * {@link AuthorRepository.ListAuthorBooksResult} carries an author name beside
+ * the title of the left-joined {@code books} row, which is {@code null} for an
+ * author without a book, and
+ * {@link AuthorRepository.CountBooksByAuthorResult} carries an author id beside
+ * the non-null {@code Long} number of that author's books.
  *
  * <p>Row absence is expressed by the {@code :optional} {@code FindAuthor}
  * query, which returns an empty {@link Optional}, while the {@code :one}
@@ -49,6 +53,16 @@
  * {@code JSONB} document as JSON text, which PostgreSQL returns normalized.
  * The enum is also bound as a query parameter to select books by format.
  *
+ * <p>The fourth migration adds the nullable {@code updated_at} column to
+ * {@code authors}, so {@link AuthorsRow} carries it as a
+ * {@link java.time.LocalDateTime}. The rename and the upsert set it with
+ * {@code now()} without binding a parameter for it, and the upsert targets the
+ * existing author of the conflicting id, so the {@code BIGSERIAL} key is never
+ * given an explicit new value. The filter binds its one {@code :name}
+ * parameter to both of its occurrences, so a {@code null} argument returns
+ * every author and a name returns only that author, and the read by id list
+ * binds one {@code List} as one server array, so an empty list matches no row.
+ *
  * <p>Every step is checked, so the process exits non-zero as soon as one
  * generated operation returns an unexpected result.
  *
@@ -218,6 +232,71 @@ public static void main(String[] args) throws SQLException {
                 + " " + cataloged.tags() + " " + cataloged.details()
         );
 
+        List<AuthorsRow> unfiltered = authors.filterAuthors(null);
+
+        checkEquals(2, unfiltered.size(), "FilterAuthors row count without a name");
+        checkEquals(created.id(), unfiltered.get(0).id(), "FilterAuthors first unfiltered id");
+        checkEquals(committedId, unfiltered.get(1).id(), "FilterAuthors second unfiltered id");
+
+        List<AuthorsRow> filtered = authors.filterAuthors("Ada Lovelace");
+
+        checkEquals(1, filtered.size(), "FilterAuthors row count for one name");
+        checkEquals(created.id(), filtered.get(0).id(), "FilterAuthors filtered id");
+        checkEquals("Ada Lovelace", filtered.get(0).name(), "FilterAuthors filtered name");
+
+        System.out.println(
+            "filtered: " + unfiltered.size() + " without a name, "
+                + filtered.get(0).name() + " by name"
+        );
+
+        List<AuthorsRow> byIds = authors.listAuthorsByIds(List.of(created.id(), committedId));
+
+        checkEquals(2, byIds.size(), "ListAuthorsByIds row count for two ids");
+        checkEquals(created.id(), byIds.get(0).id(), "ListAuthorsByIds first id");
+        checkEquals(committedId, byIds.get(1).id(), "ListAuthorsByIds second id");
+        checkEquals(
+            0,
+            authors.listAuthorsByIds(List.of()).size(),
+            "ListAuthorsByIds row count for an empty list"
+        );
+
+        System.out.println("by ids: " + byIds.get(0).id() + " " + byIds.get(1).id());
+
+        List<AuthorRepository.CountBooksByAuthorResult> bookCounts = authors.countBooksByAuthor();
+
+        checkEquals(1, bookCounts.size(), "CountBooksByAuthor row count");
+        checkEquals(committedId, bookCounts.get(0).authorId(), "CountBooksByAuthor author id");
+        checkEquals(2L, bookCounts.get(0).books(), "CountBooksByAuthor count");
+
+        System.out.println(
+            "book counts: " + bookCounts.get(0).authorId() + " / " + bookCounts.get(0).books()
+        );
+
+        int renamedRows = authors.renameAuthor("Ada Byron", created.id());
+
+        checkEquals(1, renamedRows, "RenameAuthor affected rows");
+
+        AuthorsRow renamed = authors.getAuthor(created.id());
+
+        checkEquals("Ada Byron", renamed.name(), "name after rename");
+        check(renamed.updatedAt() != null, "RenameAuthor left updatedAt null");
+
+        System.out.println("renamed: " + renamed.name() + " at " + renamed.updatedAt());
+
+        AuthorsRow upserted = authors.upsertAuthor(
+            committedId,
+            "Grace Murray Hopper",
+            "Rear admiral"
+        );
+
+        check(upserted != null, "UpsertAuthor returned no row");
+        checkEquals(committedId, upserted.id(), "UpsertAuthor id");
+        checkEquals("Grace Murray Hopper", upserted.name(), "UpsertAuthor name");
+        checkEquals("Rear admiral", upserted.bio(), "UpsertAuthor bio");
+        check(upserted.updatedAt() != null, "UpsertAuthor left updatedAt null");
+
+        System.out.println("upserted: " + upserted.name() + " / " + upserted.bio());
+
         int deletedRows = authors.deleteAuthor(created.id());
 
         checkEquals(1, deletedRows, "DeleteAuthor affected rows");

scripts/verify-release.sh

@@ -20,9 +20,12 @@
 # generates. The final generation uses the sample's own 'sqlcj.yaml' from the
 # sample directory and must restore the first generated output byte for byte
 # before the sample is compiled and run. The sample's migrations declare an
-# enum type and add an enum, an array, and a 'JSONB' column to 'books', so the
-# generated package holds the shared 'AuthorsRow' and 'BooksRow' records, the
-# 'BookFormat' enum, and one repository.
+# enum type and add an enum, an array, and a 'JSONB' column to 'books' and an
+# 'updated_at' column to 'authors', so the generated package holds the shared
+# 'AuthorsRow' and 'BooksRow' records, the 'BookFormat' enum, and one
+# repository whose methods cover every named query, including the named
+# parameter filter, the 'now()' rename, the grouped book count, the upsert, and
+# the id-list read.
 #
 # Requirements: JDK 21, Maven, psql, and a reachable PostgreSQL server.
 #
@@ -246,14 +249,18 @@ generated_count=$(find "${generated_root}" -type f -name '*.java' | wc -l)
 
 for method in createAuthor getAuthor findAuthor listAuthors updateAuthorBio deleteAuthor \
     searchAuthors countAuthors listAuthorPage createBook listAuthorBooks \
-    createCatalogedBook listBooksByFormat; do
+    createCatalogedBook listBooksByFormat filterAuthors renameAuthor \
+    countBooksByAuthor upsertAuthor listAuthorsByIds; do
     grep -q " ${method}(" "${generated_package}/AuthorRepository.java" \
         || fail "the generated repository is missing the ${method} method"
 done
 
 grep -q "LocalDateTime createdAt" "${generated_package}/AuthorsRow.java" \
     || fail "the generated row is missing the column the second migration adds"
 
+grep -q "LocalDateTime updatedAt" "${generated_package}/AuthorsRow.java" \
+    || fail "the generated row is missing the column the fourth migration adds"
+
 diff -r "${first_generation}" "${generated_root}" \
     || fail "generating from the sample directory did not restore its first generated output"
 
@@ -287,6 +294,7 @@ PGPASSWORD="${db_password}" psql \
     --file "${sample_dir}/sql/migrations/V1__create_authors_and_books.sql" \
     --file "${sample_dir}/sql/migrations/V2__add_author_created_at_and_book_index.sql" \
     --file "${sample_dir}/sql/migrations/V3__add_book_format_tags_and_details.sql" \
+    --file "${sample_dir}/sql/migrations/V4__add_author_updated_at.sql" \
     "${jdbc_url#jdbc:}"
 
 log "Compiling the sample against the staged runtime"