Add split_field() Fix function. (#100)

metafacture · Jan 11, 2022 · 1900427 · 1900427
1 parent 0155cee
commit 1900427
Show file tree

Hide file tree

Showing 2 changed files with 98 additions and 0 deletions.
diff --git a/metafix/src/main/java/org/metafacture/metafix/FixMethod.java b/metafix/src/main/java/org/metafacture/metafix/FixMethod.java
@@ -29,6 +29,7 @@
 import java.util.Set;
 import java.util.function.Function;
 import java.util.function.Predicate;
+import java.util.function.UnaryOperator;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 import java.util.stream.Collectors;
@@ -365,6 +366,21 @@ public void apply(final Metafix metafix, final Record record, final List<String>
             });
         }
     },
+    split_field {
+        public void apply(final Metafix metafix, final Record record, final List<String> params, final Map<String, String> options) {
+            record.transformField(params.get(0), v -> {
+                final String splitChar = params.size() > 1 ? params.get(1) : "\\s+";
+                final Pattern splitPattern = Pattern.compile(splitChar);
+
+                final UnaryOperator<Value> splitOperator = s ->
+                    newArray(Arrays.stream(splitPattern.split(s.asString())).map(Value::new));
+
+                return v.isString() ? splitOperator.apply(v) : v.isArray() ?
+                    newArray(v.asArray().stream().map(splitOperator)) : v.isHash() ?
+                    Value.newHash(h -> v.asHash().forEach((f, s) -> h.put(f, splitOperator.apply(s)))) : null;
+            });
+        }
+    },
     substring {
         public void apply(final Metafix metafix, final Record record, final List<String> params, final Map<String, String> options) {
             record.transformFields(params, s -> s.substring(getInteger(params, 1), getInteger(params, 2) - 1));

diff --git a/metafix/src/test/java/org/metafacture/metafix/MetafixMethodTest.java b/metafix/src/test/java/org/metafacture/metafix/MetafixMethodTest.java
@@ -850,6 +850,88 @@ public void shouldSortFieldAndRemoveDuplicates() {
         );
     }
 
+    @Test
+    public void shouldSplitStringField() {
+        MetafixTestHelpers.assertFix(streamReceiver, Arrays.asList(
+                "split_field(date, '-')"
+            ),
+            i -> {
+                i.startRecord("1");
+                i.literal("date", "1918-17-16");
+                i.endRecord();
+            },
+            o -> {
+                o.get().startRecord("1");
+                o.get().startEntity("date");
+                o.get().literal("1", "1918");
+                o.get().literal("2", "17");
+                o.get().literal("3", "16");
+                o.get().endEntity();
+                o.get().endRecord();
+            }
+        );
+    }
+
+    @Test
+    public void shouldSplitArrayField() {
+        MetafixTestHelpers.assertFix(streamReceiver, Arrays.asList(
+                "split_field(date, '-')"
+            ),
+            i -> {
+                i.startRecord("1");
+                i.literal("date", "1918-17-16");
+                i.literal("date", "2021-22-23");
+                i.endRecord();
+            },
+            (o, f) -> {
+                o.get().startRecord("1");
+                o.get().startEntity("date");
+                o.get().startEntity("1");
+                o.get().literal("1", "1918");
+                o.get().literal("2", "17");
+                o.get().literal("3", "16");
+                o.get().endEntity();
+                o.get().startEntity("2");
+                o.get().literal("1", "2021");
+                o.get().literal("2", "22");
+                o.get().literal("3", "23");
+                f.apply(2).endEntity();
+                o.get().endRecord();
+            }
+        );
+    }
+
+    @Test
+    public void shouldSplitHashField() {
+        MetafixTestHelpers.assertFix(streamReceiver, Arrays.asList(
+                "split_field(date, '-')"
+            ),
+            i -> {
+                i.startRecord("1");
+                i.startEntity("date");
+                i.literal("a", "1918-17-16");
+                i.literal("b", "2021-22-23");
+                i.endEntity();
+                i.endRecord();
+            },
+            (o, f) -> {
+                o.get().startRecord("1");
+                o.get().startEntity("date");
+                o.get().startEntity("a");
+                o.get().literal("1", "1918");
+                o.get().literal("2", "17");
+                o.get().literal("3", "16");
+                o.get().endEntity();
+                o.get().startEntity("b");
+                o.get().literal("1", "2021");
+                o.get().literal("2", "22");
+                o.get().literal("3", "23");
+                f.apply(2).endEntity();
+                o.get().endRecord();
+            }
+        );
+    }
+
     private void assertThrows(final Class<?> expectedClass, final String expectedMessage, final Executable executable) {
         final Throwable exception = Assertions.assertThrows(MetafactureException.class, executable).getCause();
         Assertions.assertSame(expectedClass, exception.getClass());