SpannerScanner: add option to disableDataboost

Allows Databoost to be disabled; it is on by default
given the point of this connector. However, there is
something to be said about compatibility so that by
default most users who haven't enabled Databoost can
still use it, but that's to be discussed for later.

Fixes #68

spark-3.1-spanner-lib/src/main/java/com/google/cloud/spark/spanner/SpannerScanner.java

@@ -89,14 +89,21 @@ public InputPartition[] planInputPartitions() {
     if (filters.length > 0) {
       sqlStmt += " WHERE " + SparkFilterUtils.getCompiledFilter(true, filters);
     }
+
+    // By default, dataBoost is enabled, given the point of this
+    // integration was to take advantage of dataBoost firstly.
+    // Please see https://github.com/GoogleCloudDataproc/spark-spanner-connector/issues/68
+    boolean disableDataboost = this.opts.get("disableDataboost");
+    boolean enableDataboost = disableDataboost == null || !disableDataboost;
+
     try (BatchReadOnlyTransaction txn =
         batchClient.batchClient.batchReadOnlyTransaction(TimestampBound.strong())) {
       String mapAsJSON = SpannerUtils.serializeMap(this.opts);
       List<com.google.cloud.spanner.Partition> rawPartitions =
           txn.partitionQuery(
               PartitionOptions.getDefaultInstance(),
               Statement.of(sqlStmt),
-              Options.dataBoostEnabled(true));
+              Options.dataBoostEnabled(enableDataboost);
 
       List<Partition> parts =
           Streams.mapWithIndex(