embulk-parser-poi_excel 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 9eabdb80c2cf4ef399fd2d9b0ec1f255cdda7ee1
-  data.tar.gz: 3b3ec43a2141a303fd234439aeef7ce2bd1da7ae
+  metadata.gz: 17272dd82557dd34357f3407c8d4eb7658e31ed2
+  data.tar.gz: e135d924eae7de15e5da96c381511a38f2d35426
 SHA512:
-  metadata.gz: f93efdbd29139564ee732ebcae17a7ecbec39b03ab9f6a161ae7ffc172760386d21c1f26aa36b399ca286b8ec315b6cd849dda3960e353085c64cbf627738b41
-  data.tar.gz: 95822b832a14fdba92b2181f178d93abff62e2dd25daeadbc4ddb4ee37cb958ca5d649c56f20f9f1c50f410e67f9fc398af28dd1adc1337204bfb91824befb15
+  metadata.gz: 7922a700f58a63aefb895334ab5483da14e4f1c36dce2beac9868de28e8c61f88ce34487c660f804d9540b176afcf73258d12c9035867271d6a020ccfe63d217
+  data.tar.gz: 4908655995c8abe3dc4bd552defb4a8142eecd093ebd5bb0b728f06030737d764e2851a31bfa479d87d3469b61a18d8beef57cc0cd40f919c34538aefd9e056a

data/README.md

@@ -10,7 +10,7 @@ This plugin uses Apache POI.
 
 ## Configuration
 
-* **sheet**: sheet name. (string, default: `Sheet1`)
+* **sheets**: sheet name. (list of string, required)
 * **skip_header_lines**: skip rows. (integer, default: `0`)
 * **columns**: column definition. see below. (hash, required)
 
@@ -24,14 +24,14 @@ This plugin uses Apache POI.
 
 ### value
 
-* **cell_value**: value in cell.
-* **cell_formula**: formula in cell. (if cell is not formula, same `cell_value`.)
-* **cell_style**: all cell style attributes. returned json string. see **attribute_name**. (**type** required `string`)
-* **cell_font**: all cell font attributes. returned json string. see **attribute_name**. (**type** required `string`)
-* **cell_comment**: all cell comment attributes. returned json string. see **attribute_name**. (**type** required `string`)
-* **sheet_name**: sheet name.
-* **row_number**: row number(1 origin).
-* **column_number**: column number(1 origin).
+* `cell_value`: value in cell.
+* `cell_formula`: formula in cell. (if cell is not formula, same `cell_value`.)
+* `cell_style`: all cell style attributes. returned json string. see **attribute_name**. (**type** required `string`)
+* `cell_font`: all cell font attributes. returned json string. see **attribute_name**. (**type** required `string`)
+* `cell_comment`: all cell comment attributes. returned json string. see **attribute_name**. (**type** required `string`)
+* `sheet_name`: sheet name.
+* `row_number`: row number(1 origin).
+* `column_number`: column number(1 origin).
 
 ### column_number
 
@@ -87,7 +87,7 @@ in:
   type: any file input plugin type
   parser:
     type: poi_excel
-    sheet: "DQ10-orb"
+    sheets: ["DQ10-orb"]
     skip_header_lines: 1	# first row is header.
     columns:
     - {name: row, type: long, value: row_number}
@@ -102,17 +102,14 @@ if omit `value`, specified `cell_value`.
 if omit `column_number` when valus is `cell_value`, specified next column.  
 if omit `column_number` when valus is `cell_style`, specified same column.
 
-### execute
+
+## Install
 
 ```
-$ cd ~/your-workspace
-$ git clone https://github.com/hishidama/embulk-parser-poi_excel.git
-$ cd embulk-parser-poi_excel
-$ ./gradlew package
-$ cd /your-embulk-working-dir
-$ embulk run -L ~/your-workspace/embulk-parser-poi_excel config.yml
+$ embulk gem install embulk-parser-poi_excel
 ```
 
+
 ## Build
 
 ```

data/build.gradle

@@ -13,7 +13,7 @@ configurations {
     provided
 }
 
-version = "0.1.0"
+version = "0.1.1"
 
 dependencies {
     compile  "org.embulk:embulk-core:0.7.5"

data/src/main/java/org/embulk/parser/poi_excel/PoiExcelParserPlugin.java

@@ -1,6 +1,7 @@
 package org.embulk.parser.poi_excel;
 
 import java.io.IOException;
+import java.util.ArrayList;
 import java.util.List;
 
 import org.apache.poi.EncryptedDocumentException;
@@ -11,6 +12,7 @@ import org.apache.poi.ss.usermodel.Workbook;
 import org.apache.poi.ss.usermodel.WorkbookFactory;
 import org.embulk.config.Config;
 import org.embulk.config.ConfigDefault;
+import org.embulk.config.ConfigException;
 import org.embulk.config.ConfigSource;
 import org.embulk.config.Task;
 import org.embulk.config.TaskSource;
@@ -26,18 +28,28 @@ import org.embulk.spi.Schema;
 import org.embulk.spi.SchemaConfig;
 import org.embulk.spi.time.TimestampParser;
 import org.embulk.spi.util.FileInputInputStream;
+import org.slf4j.Logger;
 
 import com.google.common.base.Optional;
 import com.ibm.icu.text.MessageFormat;
 
 public class PoiExcelParserPlugin implements ParserPlugin {
+	private final Logger log = Exec.getLogger(getClass());
 
 	public static final String TYPE = "poi_excel";
 
 	public interface PluginTask extends Task, TimestampParser.Task {
 		@Config("sheet")
-		@ConfigDefault("\"Sheet1\"")
-		public String getSheet();
+		@ConfigDefault("null")
+		public Optional<String> getSheet();
+
+		@Config("sheets")
+		@ConfigDefault("[]")
+		public List<String> getSheets();
+
+		@Config("ignore_sheet_not_found")
+		@ConfigDefault("false")
+		public boolean getIgnoreSheetNotFound();
 
 		@Config("skip_header_lines")
 		@ConfigDefault("0")
@@ -145,6 +157,16 @@ public class PoiExcelParserPlugin implements ParserPlugin {
 	public void run(TaskSource taskSource, Schema schema, FileInput input, PageOutput output) {
 		PluginTask task = taskSource.loadTask(PluginTask.class);
 
+		List<String> sheetNames = new ArrayList<>();
+		Optional<String> sheetOption = task.getSheet();
+		if (sheetOption.isPresent()) {
+			sheetNames.add(sheetOption.get());
+		}
+		sheetNames.addAll(task.getSheets());
+		if (sheetNames.isEmpty()) {
+			throw new ConfigException("Attribute sheets is required but not set");
+		}
+
 		try (FileInputInputStream is = new FileInputInputStream(input)) {
 			while (is.nextFile()) {
 				Workbook workbook;
@@ -154,39 +176,47 @@ public class PoiExcelParserPlugin implements ParserPlugin {
 					throw new RuntimeException(e);
 				}
 
-				String sheetName = task.getSheet();
-				Sheet sheet = workbook.getSheet(sheetName);
-				if (sheet == null) {
-					throw new RuntimeException(MessageFormat.format("not found sheet={0}", sheetName));
-				}
-
-				run(task, schema, sheet, output);
+				run(task, schema, workbook, sheetNames, output);
 			}
 		}
 	}
 
-	protected void run(PluginTask task, Schema schema, Sheet sheet, PageOutput output) {
+	protected void run(PluginTask task, Schema schema, Workbook workbook, List<String> sheetNames, PageOutput output) {
 		int skipHeaderLines = task.getSkipHeaderLines();
 		final int flushCount = task.getFlushCount();
 
-		try (final PageBuilder pageBuilder = new PageBuilder(Exec.getBufferAllocator(), schema, output)) {
-			PoiExcelVisitorFactory factory = newPoiExcelVisitorFactory(task, sheet, pageBuilder);
-			PoiExcelColumnVisitor visitor = factory.getPoiExcelColumnVisitor();
-
-			int count = 0;
-			for (final Row row : sheet) {
-				if (row.getRowNum() < skipHeaderLines) {
-					continue;
+		try (PageBuilder pageBuilder = new PageBuilder(Exec.getBufferAllocator(), schema, output)) {
+			for (String sheetName : sheetNames) {
+				Sheet sheet = workbook.getSheet(sheetName);
+				if (sheet == null) {
+					if (task.getIgnoreSheetNotFound()) {
+						log.info("ignore: not found sheet={}", sheetName);
+						continue;
+					} else {
+						throw new RuntimeException(MessageFormat.format("not found sheet={0}", sheetName));
+					}
 				}
 
-				visitor.setRow(row);
-				schema.visitColumns(visitor);
-				pageBuilder.addRecord();
+				log.info("sheet={}", sheetName);
+				PoiExcelVisitorFactory factory = newPoiExcelVisitorFactory(task, sheet, pageBuilder);
+				PoiExcelColumnVisitor visitor = factory.getPoiExcelColumnVisitor();
+
+				int count = 0;
+				for (Row row : sheet) {
+					if (row.getRowNum() < skipHeaderLines) {
+						continue;
+					}
+
+					visitor.setRow(row);
+					schema.visitColumns(visitor);
+					pageBuilder.addRecord();
 
-				if (++count >= flushCount) {
-					pageBuilder.flush();
-					count = 0;
+					if (++count >= flushCount) {
+						pageBuilder.flush();
+						count = 0;
+					}
 				}
+				pageBuilder.flush();
 			}
 			pageBuilder.finish();
 		}

data/src/main/java/org/embulk/parser/poi_excel/visitor/AbstractPoiExcelCellAttributeVisitor.java

@@ -48,7 +48,7 @@ public abstract class AbstractPoiExcelCellAttributeVisitor<A> {
 	protected abstract A getAttributeSource(Column column, ColumnOptionTask option, Cell cell);
 
 	private void visitKey(Column column, ColumnOptionTask option, String key, Cell cell, A source, CellVisitor visitor) {
-		Object value = getAttributeValue(column, option, cell, source, key);
+		Object value = getAttributeValue(column, cell, source, key);
 		if (value == null) {
 			pageBuilder.setNull(column);
 		} else if (value instanceof String) {
@@ -59,6 +59,8 @@ public abstract class AbstractPoiExcelCellAttributeVisitor<A> {
 			visitor.visitCellValueBoolean(column, source, (Boolean) value);
 		} else if (value instanceof Double) {
 			visitor.visitCellValueNumeric(column, source, (Double) value);
+		} else if (value instanceof Map) {
+			visitor.visitCellValueString(column, source, convertJsonString(value));
 		} else {
 			throw new IllegalStateException(MessageFormat.format("unsupported conversion. type={0}, value={1}", value
 					.getClass().getName(), value));
@@ -70,41 +72,46 @@ public abstract class AbstractPoiExcelCellAttributeVisitor<A> {
 
 		Optional<List<String>> nameOption = option.getAttributeName();
 		if (nameOption.isPresent()) {
-			result = new LinkedHashMap<>();
-
 			List<String> list = nameOption.get();
-			for (String key : list) {
-				Object value = getAttributeValue(column, option, cell, source, key);
-				result.put(key, value);
-			}
+			result = getSpecifiedValues(column, cell, source, list);
 		} else {
-			result = new TreeMap<>();
-
-			Collection<String> keys = getAttributeSupplierMap().keySet();
-			for (String key : keys) {
-				if (acceptKey(key)) {
-					Object value = getAttributeValue(column, option, cell, source, key);
-					result.put(key, value);
-				}
-			}
+			result = getAllValues(column, cell, source);
 		}
 
-		String json;
-		try {
-			ObjectMapper mapper = new ObjectMapper();
-			json = mapper.writeValueAsString(result);
-		} catch (JsonProcessingException e) {
-			throw new RuntimeException(e);
+		String json = convertJsonString(result);
+		visitor.visitCellValueString(column, cell, json);
+	}
+
+	protected final Map<String, Object> getSpecifiedValues(Column column, Cell cell, A source, List<String> keyList) {
+		Map<String, Object> result = new LinkedHashMap<>();
+
+		for (String key : keyList) {
+			Object value = getAttributeValue(column, cell, source, key);
+			result.put(key, value);
 		}
 
-		visitor.visitCellValueString(column, cell, json);
+		return result;
+	}
+
+	protected final Map<String, Object> getAllValues(Column column, Cell cell, A source) {
+		Map<String, Object> result = new TreeMap<>();
+
+		Collection<String> keys = getAttributeSupplierMap().keySet();
+		for (String key : keys) {
+			if (acceptKey(key)) {
+				Object value = getAttributeValue(column, cell, source, key);
+				result.put(key, value);
+			}
+		}
+
+		return result;
 	}
 
 	protected boolean acceptKey(String key) {
 		return true;
 	}
 
-	private Object getAttributeValue(Column column, ColumnOptionTask option, Cell cell, A source, String key) {
+	protected final Object getAttributeValue(Column column, Cell cell, A source, String key) {
 		Map<String, AttributeSupplier<A>> map = getAttributeSupplierMap();
 		AttributeSupplier<A> supplier = map.get(key.toLowerCase());
 		if (supplier == null) {
@@ -130,4 +137,13 @@ public abstract class AbstractPoiExcelCellAttributeVisitor<A> {
 	}
 
 	protected abstract Map<String, AttributeSupplier<A>> getAttributeSupplierMap();
+
+	protected final String convertJsonString(Object result) {
+		try {
+			ObjectMapper mapper = new ObjectMapper();
+			return mapper.writeValueAsString(result);
+		} catch (JsonProcessingException e) {
+			throw new RuntimeException(e);
+		}
+	}
 }

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellCommentVisitor.java

@@ -5,6 +5,7 @@ import java.util.HashMap;
 import java.util.Map;
 
 import org.apache.poi.ss.usermodel.Cell;
+import org.apache.poi.ss.usermodel.ClientAnchor;
 import org.apache.poi.ss.usermodel.Comment;
 import org.apache.poi.ss.usermodel.RichTextString;
 import org.embulk.parser.poi_excel.PoiExcelParserPlugin.ColumnOptionTask;
@@ -21,13 +22,20 @@ public class PoiExcelCellCommentVisitor extends AbstractPoiExcelCellAttributeVis
 		return cell.getCellComment();
 	}
 
+	protected boolean acceptKey(String key) {
+		if (key.equals("client_anchor")) {
+			return false;
+		}
+		return true;
+	}
+
 	@Override
 	protected Map<String, AttributeSupplier<Comment>> getAttributeSupplierMap() {
 		return SUPPLIER_MAP;
 	}
 
-	protected static final Map<String, AttributeSupplier<Comment>> SUPPLIER_MAP;
-	static {
+	private final Map<String, AttributeSupplier<Comment>> SUPPLIER_MAP;
+	{
 		Map<String, AttributeSupplier<Comment>> map = new HashMap<>(32);
 		map.put("author", new AttributeSupplier<Comment>() {
 			@Override
@@ -60,9 +68,27 @@ public class PoiExcelCellCommentVisitor extends AbstractPoiExcelCellAttributeVis
 				return rich.getString();
 			}
 		});
-
-		// TODO getClientAnchor
-
+		map.put("client_anchor", new AttributeSupplier<Comment>() {
+			@Override
+			public Object get(Column column, Cell cell, Comment comment) {
+				return getClientAnchorValue(column, cell, comment, null);
+			}
+		});
+		for (String key : PoiExcelClientAnchorVisitor.getKeys()) {
+			map.put("client_anchor." + key, new AttributeSupplier<Comment>() {
+				@Override
+				public Object get(Column column, Cell cell, Comment comment) {
+					return getClientAnchorValue(column, cell, comment, key);
+				}
+			});
+		}
 		SUPPLIER_MAP = Collections.unmodifiableMap(map);
 	}
+
+	final Object getClientAnchorValue(Column column, Cell cell, Comment comment, String key) {
+		ClientAnchor anchor = comment.getClientAnchor();
+		PoiExcelVisitorFactory factory = visitorValue.getVisitorFactory();
+		PoiExcelClientAnchorVisitor delegator = factory.getPoiExcelClientAnchorVisitor();
+		return delegator.getClientAnchorValue(column, cell, anchor, key);
+	}
 }

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellFontVisitor.java

@@ -31,7 +31,7 @@ public class PoiExcelCellFontVisitor extends AbstractPoiExcelCellAttributeVisito
 		return SUPPLIER_MAP;
 	}
 
-	protected static final Map<String, AttributeSupplier<Font>> SUPPLIER_MAP;
+	private static final Map<String, AttributeSupplier<Font>> SUPPLIER_MAP;
 	static {
 		Map<String, AttributeSupplier<Font>> map = new HashMap<>(32);
 		map.put("font_name", new AttributeSupplier<Font>() {

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellStyleVisitor.java

@@ -35,8 +35,7 @@ public class PoiExcelCellStyleVisitor extends AbstractPoiExcelCellAttributeVisit
 		return SUPPLIER_MAP;
 	}
 
-	protected static final Map<String, AttributeSupplier<CellStyle>> SUPPLIER_MAP;
-
+	private static final Map<String, AttributeSupplier<CellStyle>> SUPPLIER_MAP;
 	static {
 		Map<String, AttributeSupplier<CellStyle>> map = new HashMap<>(32);
 		map.put("alignment", new AttributeSupplier<CellStyle>() {

data/src/main/java/org/embulk/parser/poi_excel/visitor/{PoiExcelCellVisitor.java → PoiExcelCellValueVisitor.java}

@@ -23,13 +23,13 @@ import org.slf4j.Logger;
 
 import com.google.common.base.Optional;
 
-public class PoiExcelCellVisitor {
+public class PoiExcelCellValueVisitor {
 	private final Logger log = Exec.getLogger(getClass());
 
 	protected final PoiExcelVisitorValue visitorValue;
 	protected final PageBuilder pageBuilder;
 
-	public PoiExcelCellVisitor(PoiExcelVisitorValue visitorValue) {
+	public PoiExcelCellValueVisitor(PoiExcelVisitorValue visitorValue) {
 		this.visitorValue = visitorValue;
 		this.pageBuilder = visitorValue.getPageBuilder();
 	}

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelClientAnchorVisitor.java

@@ -0,0 +1,100 @@
+package org.embulk.parser.poi_excel.visitor;
+
+import java.util.Collection;
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.Map;
+
+import org.apache.poi.ss.usermodel.Cell;
+import org.apache.poi.ss.usermodel.ClientAnchor;
+import org.embulk.parser.poi_excel.PoiExcelParserPlugin.ColumnOptionTask;
+import org.embulk.spi.Column;
+
+public class PoiExcelClientAnchorVisitor extends AbstractPoiExcelCellAttributeVisitor<ClientAnchor> {
+
+	public PoiExcelClientAnchorVisitor(PoiExcelVisitorValue visitorValue) {
+		super(visitorValue);
+	}
+
+	public Object getClientAnchorValue(Column column, Cell cell, ClientAnchor anchor, String key) {
+		if (key == null || key.isEmpty()) {
+			return getAllValues(column, cell, anchor);
+		}
+
+		return getAttributeValue(column, cell, anchor, key);
+	}
+
+	@Override
+	protected ClientAnchor getAttributeSource(Column column, ColumnOptionTask option, Cell cell) {
+		throw new UnsupportedOperationException();
+	}
+
+	@Override
+	protected Map<String, AttributeSupplier<ClientAnchor>> getAttributeSupplierMap() {
+		return SUPPLIER_MAP;
+	}
+
+	private static final Map<String, AttributeSupplier<ClientAnchor>> SUPPLIER_MAP;
+	static {
+		Map<String, AttributeSupplier<ClientAnchor>> map = new HashMap<>(16);
+		map.put("anchor_type", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getAnchorType();
+			}
+		});
+		map.put("col1", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getCol1();
+			}
+		});
+		map.put("col2", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getCol2();
+			}
+		});
+		map.put("dx1", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getDx1();
+			}
+		});
+		map.put("dx2", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getDx2();
+			}
+		});
+		map.put("dy1", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getDy1();
+			}
+		});
+		map.put("dy2", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getDy2();
+			}
+		});
+		map.put("row1", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getRow1();
+			}
+		});
+		map.put("row2", new AttributeSupplier<ClientAnchor>() {
+			@Override
+			public Object get(Column column, Cell cell, ClientAnchor anchor) {
+				return (long) anchor.getRow2();
+			}
+		});
+		SUPPLIER_MAP = Collections.unmodifiableMap(map);
+	}
+
+	public static Collection<String> getKeys() {
+		return SUPPLIER_MAP.keySet();
+	}
+}

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelColumnIndex.java

@@ -54,6 +54,8 @@ public class PoiExcelColumnIndex {
 						CellReference.convertNumToColString(index), valueType);
 				option.setColumnIndex(index);
 				indexMap.put(column.getName(), index);
+			} else {
+				log.info("column.name={} <- value_type={}", column.getName(), valueType);
 			}
 		}
 	}

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelColumnVisitor.java

@@ -125,7 +125,7 @@ public class PoiExcelColumnVisitor implements ColumnVisitor {
 	}
 
 	private void visitCellValue(Column column, ColumnOptionTask option, Cell cell, CellVisitor visitor) {
-		PoiExcelCellVisitor delegator = factory.getPoiExcelCellVisitor();
+		PoiExcelCellValueVisitor delegator = factory.getPoiExcelCellValueVisitor();
 		delegator.visitCellValue(column, option, cell, visitor);
 	}
 

data/src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelVisitorFactory.java

@@ -100,17 +100,17 @@ public class PoiExcelVisitorFactory {
 	}
 
 	// cell value/formula
-	private PoiExcelCellVisitor poiExcelCellVisitor;
+	private PoiExcelCellValueVisitor poiExcelCellValueVisitor;
 
-	public final PoiExcelCellVisitor getPoiExcelCellVisitor() {
-		if (poiExcelCellVisitor == null) {
-			poiExcelCellVisitor = newPoiExcelCellVisitor();
+	public final PoiExcelCellValueVisitor getPoiExcelCellValueVisitor() {
+		if (poiExcelCellValueVisitor == null) {
+			poiExcelCellValueVisitor = newPoiExcelCellValueVisitor();
 		}
-		return poiExcelCellVisitor;
+		return poiExcelCellValueVisitor;
 	}
 
-	protected PoiExcelCellVisitor newPoiExcelCellVisitor() {
-		return new PoiExcelCellVisitor(visitorValue);
+	protected PoiExcelCellValueVisitor newPoiExcelCellValueVisitor() {
+		return new PoiExcelCellValueVisitor(visitorValue);
 	}
 
 	// cell style
@@ -142,19 +142,33 @@ public class PoiExcelVisitorFactory {
 	}
 
 	// cell comment
-	private PoiExcelCellCommentVisitor PoiExcelCellCommentVisitor;
+	private PoiExcelCellCommentVisitor poiExcelCellCommentVisitor;
 
 	public final PoiExcelCellCommentVisitor getPoiExcelCellCommentVisitor() {
-		if (PoiExcelCellCommentVisitor == null) {
-			PoiExcelCellCommentVisitor = newPoiExcelCellCommentVisitor();
+		if (poiExcelCellCommentVisitor == null) {
+			poiExcelCellCommentVisitor = newPoiExcelCellCommentVisitor();
 		}
-		return PoiExcelCellCommentVisitor;
+		return poiExcelCellCommentVisitor;
 	}
 
 	protected PoiExcelCellCommentVisitor newPoiExcelCellCommentVisitor() {
 		return new PoiExcelCellCommentVisitor(visitorValue);
 	}
 
+	// ClientAnchor
+	private PoiExcelClientAnchorVisitor poiExcelClientAnchorVisitor;
+
+	public final PoiExcelClientAnchorVisitor getPoiExcelClientAnchorVisitor() {
+		if (poiExcelClientAnchorVisitor == null) {
+			poiExcelClientAnchorVisitor = newPoiExcelClientAnchorVisitor();
+		}
+		return poiExcelClientAnchorVisitor;
+	}
+
+	protected PoiExcelClientAnchorVisitor newPoiExcelClientAnchorVisitor() {
+		return new PoiExcelClientAnchorVisitor(visitorValue);
+	}
+
 	// color
 	private PoiExcelColorVisitor poiExcelColorVisitor;
 

data/src/test/java/org/embulk/parser/poi_excel/TestPoiExcelParserPlugin.java

@@ -25,7 +25,7 @@ public class TestPoiExcelParserPlugin {
 			tester.addParserPlugin(PoiExcelParserPlugin.TYPE, PoiExcelParserPlugin.class);
 
 			EmbulkTestParserConfig parser = tester.newParserConfig(PoiExcelParserPlugin.TYPE);
-			parser.set("sheet", "test1");
+			parser.set("sheets", Arrays.asList("test1"));
 			parser.set("skip_header_lines", 1);
 			parser.set("default_timezone", "Asia/Tokyo");
 			parser.addColumn("boolean", "boolean");
@@ -73,7 +73,7 @@ public class TestPoiExcelParserPlugin {
 			tester.addParserPlugin(PoiExcelParserPlugin.TYPE, PoiExcelParserPlugin.class);
 
 			EmbulkTestParserConfig parser = tester.newParserConfig(PoiExcelParserPlugin.TYPE);
-			parser.set("sheet", "test1");
+			parser.set("sheets", Arrays.asList("test1"));
 			parser.set("skip_header_lines", 1);
 			parser.set("cell_error_null", false);
 			parser.addColumn("sheet", "string").set("value", "sheet_name");
@@ -184,4 +184,24 @@ public class TestPoiExcelParserPlugin {
 		assertThat(r.getAsString("a"), is(a));
 		assertThat(r.getAsString("b"), is(b));
 	}
+
+	@Test
+	public void test_sheets() throws ParseException {
+		try (EmbulkPluginTester tester = new EmbulkPluginTester()) {
+			tester.addParserPlugin(PoiExcelParserPlugin.TYPE, PoiExcelParserPlugin.class);
+
+			EmbulkTestParserConfig parser = tester.newParserConfig(PoiExcelParserPlugin.TYPE);
+			parser.set("sheets", Arrays.asList("formula_replace", "merged_cell"));
+			parser.addColumn("a", "string");
+
+			URL inFile = getClass().getResource("test1.xls");
+			List<OutputRecord> result = tester.runParser(inFile, parser);
+
+			assertThat(result.size(), is(2 + 4));
+			assertThat(result.get(0).getAsString("a"), is("boolean"));
+			assertThat(result.get(1).getAsString("a"), is("test2-b1"));
+			assertThat(result.get(2).getAsString("a"), is("test3-a1"));
+			assertThat(result.get(3).getAsString("a"), is("data"));
+		}
+	}
 }

data/src/test/java/org/embulk/parser/poi_excel/TestPoiExcelParserPlugin_cellComment.java

@@ -1,10 +1,13 @@
 package org.embulk.parser.poi_excel;
 
 import static org.hamcrest.CoreMatchers.is;
+import static org.hamcrest.CoreMatchers.nullValue;
 import static org.junit.Assert.assertThat;
+import static org.junit.Assert.fail;
 
 import java.net.URL;
 import java.text.ParseException;
+import java.util.Arrays;
 import java.util.List;
 
 import org.embulk.parser.EmbulkPluginTester;
@@ -39,4 +42,58 @@ public class TestPoiExcelParserPlugin_cellComment {
 		assertThat(record.getAsString("comment"), is(comment));
 		assertThat(record.getAsString("author"), is(author));
 	}
+
+	@Test
+	public void testComment_all() throws ParseException {
+		try (EmbulkPluginTester tester = new EmbulkPluginTester()) {
+			tester.addParserPlugin(PoiExcelParserPlugin.TYPE, PoiExcelParserPlugin.class);
+
+			EmbulkTestParserConfig parser = tester.newParserConfig(PoiExcelParserPlugin.TYPE);
+			parser.set("sheet", "comment");
+			parser.addColumn("comment", "string").set("value", "cell_comment");
+
+			URL inFile = getClass().getResource("test1.xls");
+			List<OutputRecord> result = tester.runParser(inFile, parser);
+
+			assertThat(result.size(), is(2));
+			check2(result, 0, "hishidama", "hishidama:\\nmy comment");
+			check2(result, 1, null, null);
+		}
+	}
+
+	private void check2(List<OutputRecord> result, int index, String author, String comment) {
+		OutputRecord record = result.get(index);
+		// System.out.println(record);
+		String s = record.getAsString("comment");
+		if (author == null && comment == null) {
+			assertThat(s, is(nullValue()));
+			return;
+		}
+
+		if (!s.contains(String.format("\"author\":\"%s\"", author))) {
+			fail(s);
+		}
+		if (!s.contains(String.format("\"string\":\"%s\"", comment))) {
+			fail(s);
+		}
+	}
+
+	@Test
+	public void testComment_keys() throws ParseException {
+		try (EmbulkPluginTester tester = new EmbulkPluginTester()) {
+			tester.addParserPlugin(PoiExcelParserPlugin.TYPE, PoiExcelParserPlugin.class);
+
+			EmbulkTestParserConfig parser = tester.newParserConfig(PoiExcelParserPlugin.TYPE);
+			parser.set("sheet", "comment");
+			parser.addColumn("comment", "string").set("value", "cell_comment")
+					.set("attribute_name", Arrays.asList("author", "string"));
+
+			URL inFile = getClass().getResource("test1.xls");
+			List<OutputRecord> result = tester.runParser(inFile, parser);
+
+			assertThat(result.size(), is(2));
+			check2(result, 0, "hishidama", "hishidama:\\nmy comment");
+			check2(result, 1, null, null);
+		}
+	}
 }

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: embulk-parser-poi_excel
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - hishidama
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2015-10-12 00:00:00.000000000 Z
+date: 2015-10-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   requirement: !ruby/object:Gem::Requirement
@@ -61,7 +61,8 @@ files:
 - src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellCommentVisitor.java
 - src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellFontVisitor.java
 - src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellStyleVisitor.java
-- src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellVisitor.java
+- src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelCellValueVisitor.java
+- src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelClientAnchorVisitor.java
 - src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelColorVisitor.java
 - src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelColumnIndex.java
 - src/main/java/org/embulk/parser/poi_excel/visitor/PoiExcelColumnVisitor.java
@@ -84,7 +85,7 @@ files:
 - src/test/java/org/embulk/parser/poi_excel/TestPoiExcelParserPlugin_columnNumber.java
 - src/test/resources/org/embulk/parser/poi_excel/test1.xls
 - classpath/commons-codec-1.9.jar
-- classpath/embulk-parser-poi_excel-0.1.0.jar
+- classpath/embulk-parser-poi_excel-0.1.1.jar
 - classpath/embulk-standards-0.7.5.jar
 - classpath/poi-3.13.jar
 - classpath/poi-ooxml-3.13.jar