Merge remote-tracking branch 'remotes/origin/main' into feature/match…

…_by_column_name
trocco-io · Mar 26, 2024 · 19216a7 · 19216a7
2 parents 2d4b9bf + 9cfaa02
commit 19216a7
Show file tree

Hide file tree

Showing 3 changed files with 42 additions and 29 deletions.
diff --git a/src/main/java/org/embulk/output/SnowflakeOutputPlugin.java b/src/main/java/org/embulk/output/SnowflakeOutputPlugin.java
@@ -11,6 +11,7 @@
 import net.snowflake.client.jdbc.internal.org.bouncycastle.pkcs.PKCSException;
 import org.embulk.config.ConfigDiff;
 import org.embulk.config.ConfigException;
+import org.embulk.config.ConfigSource;
 import org.embulk.config.TaskSource;
 import org.embulk.output.jdbc.*;
 import org.embulk.output.snowflake.PrivateKeyReader;
@@ -27,8 +28,6 @@
 import org.embulk.util.config.ConfigDefault;
 
 public class SnowflakeOutputPlugin extends AbstractJdbcOutputPlugin {
-  private StageIdentifier stageIdentifier;
-
   public interface SnowflakePluginTask extends PluginTask {
     @Config("driver_path")
     @ConfigDefault("null")
@@ -79,6 +78,10 @@ public interface SnowflakePluginTask extends PluginTask {
     @ConfigDefault("true")
     public boolean getEmtpyFieldAsNull();
 
+    @Config("delete_stage_on_error")
+    @ConfigDefault("false")
+    public boolean getDeleteStageOnError();
+
     @Config("match_by_column_name")
     @ConfigDefault("\"none\"")
     public MatchByColumnName getMatchByColumnName();
@@ -188,25 +191,39 @@ protected JdbcOutputConnector getConnector(PluginTask task, boolean retryableMet
   }
 
   @Override
-  public ConfigDiff resume(
-      TaskSource taskSource, Schema schema, int taskCount, OutputPlugin.Control control) {
-    throw new UnsupportedOperationException("snowflake output plugin does not support resuming");
-  }
-
-  @Override
-  protected void doCommit(JdbcOutputConnection con, PluginTask task, int taskCount)
-      throws SQLException {
-    super.doCommit(con, task, taskCount);
-    SnowflakeOutputConnection snowflakeCon = (SnowflakeOutputConnection) con;
-
+  public ConfigDiff transaction(
+      ConfigSource config, Schema schema, int taskCount, OutputPlugin.Control control) {
+    PluginTask task = CONFIG_MAPPER.map(config, this.getTaskClass());
     SnowflakePluginTask t = (SnowflakePluginTask) task;
-    if (this.stageIdentifier == null) {
-      this.stageIdentifier = StageIdentifierHolder.getStageIdentifier(t);
+    StageIdentifier stageIdentifier = StageIdentifierHolder.getStageIdentifier(t);
+    ConfigDiff configDiff;
+    SnowflakeOutputConnection snowflakeCon = null;
+
+    try {
+      snowflakeCon = (SnowflakeOutputConnection) getConnector(task, true).connect(true);
+      snowflakeCon.runCreateStage(stageIdentifier);
+      configDiff = super.transaction(config, schema, taskCount, control);
+      if (t.getDeleteStage()) {
+        snowflakeCon.runDropStage(stageIdentifier);
+      }
+    } catch (Exception e) {
+      if (t.getDeleteStage() && t.getDeleteStageOnError()) {
+        try {
+          snowflakeCon.runDropStage(stageIdentifier);
+        } catch (SQLException ex) {
+          throw new RuntimeException(ex);
+        }
+      }
+      throw new RuntimeException(e);
     }
 
-    if (t.getDeleteStage()) {
-      snowflakeCon.runDropStage(this.stageIdentifier);
-    }
+    return configDiff;
+  }
+
+  @Override
+  public ConfigDiff resume(
+      TaskSource taskSource, Schema schema, int taskCount, OutputPlugin.Control control) {
+    throw new UnsupportedOperationException("snowflake output plugin does not support resuming");
   }
 
   @Override
@@ -255,20 +272,11 @@ protected BatchInsert newBatchInsert(PluginTask task, Optional<MergeConfig> merg
       throw new UnsupportedOperationException(
           "Snowflake output plugin doesn't support 'merge_direct' mode.");
     }
-
-    SnowflakePluginTask t = (SnowflakePluginTask) task;
-    // TODO: put some where executes once
-    if (this.stageIdentifier == null) {
-      SnowflakeOutputConnection snowflakeCon =
-          (SnowflakeOutputConnection) getConnector(task, true).connect(true);
-      this.stageIdentifier = StageIdentifierHolder.getStageIdentifier(t);
-      snowflakeCon.runCreateStage(this.stageIdentifier);
-    }
     SnowflakePluginTask pluginTask = (SnowflakePluginTask) task;
 
     return new SnowflakeCopyBatchInsert(
         getConnector(task, true),
-        this.stageIdentifier,
+        StageIdentifierHolder.getStageIdentifier(pluginTask),
         pluginTask.getCopyIntoTableColumnNames(),
         pluginTask.getCopyIntoCSVColumnNumbers(),
         false,

diff --git a/src/main/java/org/embulk/output/snowflake/SnowflakeCopyBatchInsert.java b/src/main/java/org/embulk/output/snowflake/SnowflakeCopyBatchInsert.java
@@ -69,7 +69,6 @@ public SnowflakeCopyBatchInsert(
   @Override
   public void prepare(TableIdentifier loadTable, JdbcSchema insertSchema) throws SQLException {
     this.connection = (SnowflakeOutputConnection) connector.connect(true);
-    this.connection.runCreateStage(stageIdentifier);
     this.tableIdentifier = loadTable;
   }
 

diff --git a/src/main/java/org/embulk/output/snowflake/SnowflakeOutputConnection.java b/src/main/java/org/embulk/output/snowflake/SnowflakeOutputConnection.java
@@ -11,8 +11,12 @@
 import org.embulk.output.jdbc.JdbcSchema;
 import org.embulk.output.jdbc.MergeConfig;
 import org.embulk.output.jdbc.TableIdentifier;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
 
 public class SnowflakeOutputConnection extends JdbcOutputConnection {
+  private final Logger logger = LoggerFactory.getLogger(SnowflakeOutputConnection.class);
+
   public SnowflakeOutputConnection(Connection connection) throws SQLException {
     super(connection, null);
   }
@@ -45,11 +49,13 @@ public void runCopy(
   public void runCreateStage(StageIdentifier stageIdentifier) throws SQLException {
     String sql = buildCreateStageSQL(stageIdentifier);
     runUpdate(sql);
+    logger.info("SQL: {}", sql);
   }
 
   public void runDropStage(StageIdentifier stageIdentifier) throws SQLException {
     String sql = buildDropStageSQL(stageIdentifier);
     runUpdate(sql);
+    logger.info("SQL: {}", sql);
   }
 
   public void runUploadFile(