Merge pull request #32 from miraisolutions/feature/30_sparse_bit_set

Use SparseBitSet instead of java.util.BitSet
miraisolutions · Feb 12, 2024 · 545829d · 545829d
2 parents 80aa717 + af6430a
commit 545829d
Show file tree

Hide file tree

Showing 4 changed files with 17 additions and 14 deletions.
diff --git a/pom.xml b/pom.xml
@@ -21,7 +21,7 @@
     <groupId>com.miraisolutions</groupId>
     <artifactId>XLConnect</artifactId>
     <packaging>jar</packaging>
-    <version>2.0.0-SNAPSHOT</version>
+    <version>2.0.1-SNAPSHOT</version>
     <name>XLConnect</name>
     <url>https://mirai-solutions.ch/</url>
     <repositories>

diff --git a/src/main/java/com/miraisolutions/xlconnect/data/Column.java b/src/main/java/com/miraisolutions/xlconnect/data/Column.java
@@ -19,16 +19,16 @@
  */
 package com.miraisolutions.xlconnect.data;
 
-import java.util.BitSet;
+import com.zaxxer.sparsebits.SparseBitSet;
 import java.util.Date;
 
 public final class Column {
     private final Object data;
     private final int size;
-    private final BitSet missing;
+    private final SparseBitSet missing;
     private final DataType type;
 
-    public Column(Object data, int size, BitSet missing, DataType type) {
+    public Column(Object data, int size, SparseBitSet missing, DataType type) {
         this.data = data;
         this.size = size;
         this.missing = missing;
@@ -55,7 +55,7 @@ public String[] getStringData() {
         return (String[]) data;
     }
 
-    public BitSet getMissing() {
+    public SparseBitSet getMissing() {
         return missing;
     }
 

diff --git a/src/main/java/com/miraisolutions/xlconnect/data/ColumnBuilder.java b/src/main/java/com/miraisolutions/xlconnect/data/ColumnBuilder.java
@@ -22,6 +22,7 @@
 import com.miraisolutions.xlconnect.ErrorBehavior;
 import com.miraisolutions.xlconnect.Workbook;
 import com.miraisolutions.xlconnect.utils.CellUtils;
+import com.zaxxer.sparsebits.SparseBitSet;
 import org.apache.poi.ss.usermodel.*;
 
 import java.util.*;
@@ -105,7 +106,7 @@ protected void addValue(Cell c, CellValue cv, DataType dt) {
     public Column buildBooleanColumn() {
         int size = values.size();
         boolean[] colValues = new boolean[size];
-        BitSet missing = new BitSet(size);
+        SparseBitSet missing = new SparseBitSet(size);
 
         int counter = 0;
         for (CellValue cv : values) {
@@ -140,7 +141,7 @@ public Column buildBooleanColumn() {
     public Column buildDateTimeColumn() {
         int size = values.size();
         Date[] colValues = new Date[size];
-        BitSet missing = new BitSet(size);
+        SparseBitSet missing = new SparseBitSet(size);
 
         Iterator<CellValue> it = values.iterator();
         Iterator<Cell> jt = cells.iterator();
@@ -196,7 +197,7 @@ public Column buildDateTimeColumn() {
     public Column buildNumericColumn() {
         int size = values.size();
         double[] colValues = new double[size];
-        BitSet missing = new BitSet(size);
+        SparseBitSet missing = new SparseBitSet(size);
 
         int counter = 0;
         for (CellValue cv : values) {
@@ -242,7 +243,7 @@ public Column buildNumericColumn() {
     public Column buildStringColumn() {
         int size = values.size();
         String[] colValues = new String[size];
-        BitSet missing = new BitSet(size);
+        SparseBitSet missing = new SparseBitSet(size);
 
         Iterator<CellValue> it = values.iterator();
         Iterator<Cell> jt = cells.iterator();

diff --git a/src/main/java/com/miraisolutions/xlconnect/integration/r/RDataFrameWrapper.java b/src/main/java/com/miraisolutions/xlconnect/integration/r/RDataFrameWrapper.java
@@ -25,7 +25,7 @@
 import com.miraisolutions.xlconnect.data.DataType;
 
 import java.util.Arrays;
-import java.util.BitSet;
+import com.zaxxer.sparsebits.SparseBitSet;
 import java.util.Date;
 import java.util.stream.IntStream;
 
@@ -90,14 +90,16 @@ public long[] getDateTimeColumn(int col) {
 
     public boolean[] isMissing(int col) {
         Column column = dataFrame.getColumn(col);
-        BitSet missing = column.getMissing();
+        SparseBitSet missing = column.getMissing();
         boolean[] na = new boolean[column.size()];
-        missing.stream().forEach(i -> na[i] = true);
+        for (int i = missing.nextSetBit(0); i >= 0; i = missing.nextSetBit(i + 1)) {
+            na[i] = true;
+        }
         return na;
     }
 
-    private static BitSet toBitSet(boolean[] bits) {
-        BitSet bs = new BitSet(bits.length);
+    private static SparseBitSet toBitSet(boolean[] bits) {
+        SparseBitSet bs = new SparseBitSet(bits.length);
         for (int i = 0; i < bits.length; i++) {
             bs.set(i, bits[i]);
         }