asfimport
commented
3 years ago Ranga Reddy: While checking createString() method implementations in FlatBufferBuilder.java class, it has two methods one is accepting CharSequence and another one is accepting the ByteBuffer as a argument.
public int createString(CharSequence s) {
}
public int createString(ByteBuffer s) {
}While checking getField() method implementation in Field.java class, here it is passed String value.
public class Field {
private final String name;
public int getField(FlatBufferBuilder builder)
{
int nameOffset = name == null ? -1 : builder.createString(name);
}
}To fix this issue, we need to pass either CharSequence or ByteBuffer as the argument in getField() method.
Solution:
public int getField(FlatBufferBuilder builder) {
java.nio.ByteBuffer bb = java.nio.ByteBuffer.wrap(name.getBytes());
int nameOffset = name == null ? -1 : builder.createString(bb);
.......
}
While running pyspark3 with pandas 1.1.5 and pyarrow 2.0.0 getting the below error:
Spark Code:
Exception:
Reporter: Ranga Reddy
Note: This issue was originally created as ARROW-14038. Please see the migration documentation for further details.