@Override
public void addBinary(Binary value) {
holder.buffer = buf = buf.reallocIfNeeded(value.length());
buf.setBytes(0, value.toByteBuffer());
holder.start = 0;
holder.end = value.length();
writer.write(holder);
}
public static VarCharHolder getVarCharHolder(BufferAllocator a, String s) {
VarCharHolder vch = new VarCharHolder();
byte[] b = s.getBytes(Charsets.UTF_8);
vch.start = 0;
vch.end = b.length;
vch.buffer = a.buffer(b.length); //
vch.buffer.setBytes(0, b);
return vch;
}
public static VarCharHolder getVarCharHolder(DrillBuf buf, String s) {
VarCharHolder vch = new VarCharHolder();
byte[] b = s.getBytes(Charsets.UTF_8);
vch.start = 0;
vch.end = b.length;
vch.buffer = buf.reallocIfNeeded(b.length);
vch.buffer.setBytes(0, b);
return vch;
}
private void writeString(
String readString, final MapOrListWriterImpl writer, String fieldName, boolean isList) {
int length;
byte[] strBytes;
try {
strBytes = readString.getBytes("UTF-8");
} catch (UnsupportedEncodingException e) {
throw new DrillRuntimeException("Unable to read string value for field: " + fieldName, e);
}
length = strBytes.length;
ensure(length);
workBuf.setBytes(0, strBytes);
final VarCharHolder vh = new VarCharHolder();
vh.buffer = workBuf;
vh.start = 0;
vh.end = length;
if (isList == false) {
writer.varChar(fieldName).write(vh);
} else {
writer.list.varChar().write(vh);
}
}
@Override
public void eval() {
out.buffer = string.buffer;
// handle invalid values; e.g. SUBSTRING(value, 0, x) or SUBSTRING(value, x, 0)
if (offset.value == 0 || length.value <= 0) {
out.start = 0;
out.end = 0;
} else {
// start iterating over the UTF-8 buffer to find the first character of the substring
int byteCount = 0;
int charCount = 0;
int byteStart = 0;
int charStart = 0;
int byteEnd = 0;
byte currentByte;
while (byteCount < string.end - string.start) {
currentByte = string.buffer.getByte(string.start + byteCount);
// check current position matches the (positive) start position
if (offset.value > 0 && charCount == (int) offset.value - 1) {
byteStart = byteCount;
charStart = charCount;
}
// check current position matches the supplied length
if (offset.value > 0 && charCount - charStart == (int) length.value) {
byteEnd = byteCount;
break;
}
if (currentByte < 128) ++charCount;
++byteCount;
}
out.start = string.start + byteStart;
out.end = string.start + byteEnd;
// search backwards for negative offsets
if (offset.value < 0) {
int endBytePos = --byteCount;
int endCharPos = --charCount;
while (byteCount >= 0) {
currentByte = string.buffer.getByte(byteCount);
if (endCharPos - charCount == -(int) offset.value) {
// matched the negative start offset
out.start = byteCount;
charCount = 0;
// search forward until we find <length> characters
while (byteCount <= endBytePos) {
currentByte = string.buffer.getByte(byteCount);
if (currentByte < 128) ++charCount;
++byteCount;
if (charCount == (int) length.value) {
out.end = byteCount;
break;
}
}
break;
}
if (currentByte < 128) --charCount;
--byteCount;
}
}
// if length exceeds value, stop at end of value
if (out.end == 0) {
out.end = string.end;
}
}
}
@Override
public void eval() {
out.buffer = in.buffer;
out.start = in.start;
out.end = in.end;
}