Merge pull request apache#113 from sun-rui/stringHashCodeInC

Implement string hash code in C.

pkg/NAMESPACE

@@ -61,3 +61,4 @@ export(
        "setCheckpointDir"
       )
 export("sparkR.init")
+useDynLib(SparkR, stringHashCode)

pkg/R/utils.R

@@ -146,27 +146,6 @@ getDependencies <- function(name) {
   binData
 }
 
-# Helper function used to wrap a 'numeric' value to integer bounds.
-# Useful for implementing C-like integer arithmetic
-wrapInt <- function(value) {
-  if (value > .Machine$integer.max) {
-    value <- value - 2 * .Machine$integer.max - 2
-  } else if (value < -1 * .Machine$integer.max) {
-    value <- 2 * .Machine$integer.max + value + 2
-  }
-  value
-}
-
-# Multiply `val` by 31 and add `addVal` to the result. Ensures that
-# integer-overflows are handled at every step.
-mult31AndAdd <- function(val, addVal) {
-  vec <- c(bitwShiftL(val, c(4,3,2,1,0)), addVal)
-  Reduce(function(a, b) {
-          wrapInt(as.numeric(a) + as.numeric(b))
-         },
-         vec)
-}
-
 #' Compute the hashCode of an object
 #'
 #' Java-style function to compute the hashCode for the given object. Returns
@@ -191,17 +170,7 @@ hashCode <- function(key) {
     intBits <- packBits(rawToBits(rawVec), "integer")
     as.integer(bitwXor(intBits[2], intBits[1]))
   } else if (class(key) == "character") {
-    n <- nchar(key)
-    if (n == 0) {
-      0L
-    } else {
-      asciiVals <- sapply(charToRaw(key), function(x) { strtoi(x, 16L) })
-      hashC <- 0
-      for (k in 1:length(asciiVals)) {
-        hashC <- mult31AndAdd(hashC, asciiVals[k])
-      }
-      as.integer(hashC)
-    }
+    .Call("stringHashCode", key)
   } else {
     warning(paste("Could not hash object, returning 0", sep=""))
     as.integer(0)

pkg/src/Makefile

@@ -31,7 +31,7 @@ else
 	MAVEN_YARN_FLAG := 
 endif
 
-all: $(TARGET_NAME)
+all: $(TARGET_NAME) sharelib
 
 $(SBT_TARGET_NAME): build.sbt $(SCALA_FILES) $(RESOURCE_FILES)
 	./sbt/sbt assembly
@@ -41,12 +41,17 @@ $(MAVEN_TARGET_NAME): pom.xml $(SCALA_FILES) $(RESOURCE_FILES)
 	mvn -Dhadoop.version=$(SPARK_HADOOP_VERSION) -Dspark.version=$(SPARK_VERSION) -DskipTests $(MAVEN_YARN_FLAG) -Dyarn.version=$(SPARK_YARN_VERSION) clean package shade:shade
 	cp -f $(MAVEN_TARGET_NAME) ../inst/$(JAR_NAME)
 
+sharelib: string_hash_code.c
+	R CMD SHLIB -o SparkR.so string_hash_code.c
+
 clean:
 	$(BUILD_TOOL) clean
 	rm -rf target
 	rm -rf project/target
 	rm -rf project/project
 	-rm sbt/sbt-launch-*.jar
-	rm -f ../inst/$(JAR_NAME)
-
+	rm -f ../inst/$(JAR_NAME) 
+	rm -f *.o
+	rm -f *.so
+
 .PHONY: all clean

pkg/src/Makefile.win

@@ -18,18 +18,23 @@ SPARK_YARN_VERSION ?= 2.4.0
 
 TARGET_NAME := $(MAVEN_TARGET_NAME)
 
-all: $(TARGET_NAME)
+all: $(TARGET_NAME) sharelib
 
 $(MAVEN_TARGET_NAME): $(SCALA_FILES) $(RESOURCE_FILES)
 	mvn.bat -Dhadoop.version=$(SPARK_HADOOP_VERSION) -Dspark.version=$(SPARK_VERSION) -Dyarn.version=$(SPARK_YARN_VERSION) -DskipTests clean package shade:shade
 	cp -f $(MAVEN_TARGET_NAME) ../inst/$(JAR_NAME)
 
+sharelib: string_hash_code.c
+	R CMD SHLIB -o SparkR.dll string_hash_code.c
+
 clean:
 	mvn.bat clean
 	rm -rf target
 	rm -rf project/target
 	rm -rf project/project
 	-rm sbt/sbt-launch-*.jar
-	rm -f ../inst/$(JAR_NAME)
-
+	rm -f ../inst/$(JAR_NAME) 
+	rm -f *.o
+	rm -f *.dll
+
 .PHONY: all clean

pkg/src/string_hash_code.c

@@ -0,0 +1,27 @@
+/*
+ * A C function for R extension which implements the Java String hash algorithm.
+ * Refer to http://en.wikipedia.org/wiki/Java_hashCode%28%29#The_java.lang.String_hash_function
+ *
+ */
+
+#include <R.h>
+#include <Rinternals.h>
+
+SEXP stringHashCode(SEXP string) {
+  const char* str;
+  R_xlen_t len, i;
+  int hashCode = 0;
+
+  if (!IS_SCALAR(string, STRSXP)) {
+    error("invalid input");
+  }
+
+  str = CHAR(asChar(string));
+  len = XLENGTH(asChar(string));
+
+  for (i = 0; i < len; i++) {
+    hashCode = (hashCode << 5) - hashCode + *str++;
+  }
+
+  return ScalarInteger(hashCode);
+}