sadikovi commented on a change in pull request #34611:
URL: https://github.com/apache/spark/pull/34611#discussion_r754794034
##########
File path:
sql/core/src/main/java/org/apache/spark/sql/execution/datasources/parquet/VectorizedPlainValuesReader.java
##########
@@ -53,19 +53,47 @@ public void skip() {
throw new UnsupportedOperationException();
}
+ private void updateCurrentByte() {
+ try {
+ currentByte = (byte) in.read();
+ } catch (IOException e) {
+ throw new ParquetDecodingException("Failed to read a byte", e);
+ }
+ }
+
@Override
public final void readBooleans(int total, WritableColumnVector c, int rowId)
{
- // TODO: properly vectorize this
- for (int i = 0; i < total; i++) {
- c.putBoolean(rowId + i, readBoolean());
+ int i = 0;
+ if (bitOffset > 0) {
+ i = Math.min(8 - bitOffset, total);
+ c.putBooleans(rowId, i, currentByte, bitOffset);
+ bitOffset = (bitOffset + i) & 7;
+ }
+ for (; i + 7 < total; i += 8) {
+ updateCurrentByte();
+ c.putBooleans(rowId + i, currentByte);
+ }
+ if (i < total) {
+ updateCurrentByte();
+ bitOffset = total - i;
+ c.putBooleans(rowId + i, bitOffset, currentByte, 0);
}
}
@Override
public final void skipBooleans(int total) {
- // TODO: properly vectorize this
- for (int i = 0; i < total; i++) {
- readBoolean();
+ // using >>3 instead of /8 below since Java division rounds towards zero
i.e. (-1)/8=0
Review comment:
I should have clarified my comment - total should be larger than 8 to
skip at least 1 byte. Thanks for pointing it out.
What I meant was that comment does not reflect the actual purpose of bit
shift and is confusing. All you are trying to figure out is whether
numBytesToSkip is positive or not so both approaches would work just fine. Bit
shift tends to be faster than the division operation - I am not sure if JVM is
smart enough to realise that these operations are equivalent in this case.
By the way, shouldn't
```
if (bitOffset > 0) {
updateCurrentByte();
}
```
be outside of the `if (numBytesToSkip >= 0) {`? In other words, what happens
when the total is 9 and bitOffset is 0?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
