decode Java int array in python code -- found a workaround for my case

Question

I get some data from cassandra's jmx. one element is defined like this. https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/db/ColumnFamilyStoreMBean.java#L358

The returned type in Java is int[];

import xml.etree.ElementTree as xml
text = requests.get('http://:8081/mbean?objectname=org.apache.cassandra.db:type=xxx,keyspace=x,columnfamily=x&template=identity').text
x = xml.fromstring(text)
>>> for i in x:
...     if i.get('name') == 'SSTableCountPerLevel':
...             p=i

>>> p.items()
[('description', 'Attribute exposed for management'), ('aggregation', 'array'), ('value', '[I@48a6ea00'), ('isnull', 'false'), ('strinit', 'false'), ('type', '[I'), ('availability', 'RO'), ('name', 'SSTableCountPerLevel')]
>>> p.get('value')
'[I@48a6ea00'

So How can I convert this [I@48a6ea00 to a list of int in python?

@update 1:

Seems JPY is here for this purpose, but since the lack of document, I will try this and give a update soon.

https://github.com/bcdev/jpy

>>> import jpyutil
>>> jpyutil.init_jvm(jvm_maxmem='512M')
>>> import jpy
>>> b = jpy.array('int', [1,2,3])
>>> b.toString()
'[I@3f2a09d5'
>>> b
[I(objectRef=0x7fa66340dfe8)
>>> b[0]
1L
>>> b[2]
3L
>>> dir(b)
['__class__', '__delattr__', '__delitem__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__getitem__', '__gt__', '__hash__', '__init__', '__jinit__', '__le__', '__len__', '__lt__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__setitem__', '__sizeof__', '__str__', '__subclasshook__', 'equals', 'getClass', 'hashCode', 'notify', 'notifyAll', 'toString', 'wait']

But I still can't convert string to Python List, which is I want.

@ conclusion:

Decode Java int array in python code should be possible if the python knows the content of java int array, but in the above case, [I@48a6ea00 is not the content of java int array, it is a memory reference to the java int array, so it is impossible.

I got this string via MX4J, I will check MX4J document to see how MX4j handle array.

@ update 2:

After checked the MX4J doucment, it is clearly stated with

Obviously it has some limitations, like not being able to manipulate data which cannot be obtained from Strings

At last, I used JYTHON, it resolves my issue perfectly.

That looks like (though might not be) the internal representation of the array by the JVM. docs.oracle.com/javase/specs/jvms/se7/jvms7.pdf (Section 4.3) might shed some light on it. Though if I'm reading it right that just says: Single Dimension Array of Integers at address 0x48a6ea00... so you might not be able to get the actual contents from that string... So as others have mentioned, you might be dealing with bad data in the first place... — LexyStardust
– LexyStardust, Commented Jan 6, 2015 at 10:29
@LexyStardust, actually I get this data via MX4J (cassandra used this Java library to support some remote call), Do you mean [I@48a6ea00 just a address of the real array? if so, MX4J should be very bad designed here, I prefer to not believe that. Thank you very much for the pdf and a new direction. — zhihuifan
– zhihuifan, Commented Jan 7, 2015 at 0:54

Dongqing · Accepted Answer · 2015-01-07 03:50:36Z

1

Looks like somewhere the value is transform incorrectly to the reference to the array ( int[] ). Could you print the original response? i.e. text , and the xml, i.e. x to check which step goes wrong.

So the problem is the original response only return the internal descriptor of the array, '[I@48a6ea00' is the reference of that int array , it's a type@address format,only makes sense inside a particular JVM, it didn't carry the actual value information. i.e, no way to get the value as it is already lost in the response.

edited Jan 7, 2015 at 3:50

answered Jan 6, 2015 at 10:20

Dongqing

6861 gold badge5 silver badges19 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

zhihuifan Over a year ago

why do you think this value is incorrectly? I have no idea who to judge. Here is the raw text. <Attribute aggregation="array" availability="RO" description="Attribute exposed for management" isnull="false" name="SSTableCountPerLevel" strinit="false" type="[I" value="[I@78a1aae0"></Attribute>. (whole message is too long to be placed here)

Dongqing Over a year ago

So the problem is there, the original response only return the internal descriptor of the array, '[I@48a6ea00' is the reference of that int array , it's a type@address format, it only makes sense inside a particular JVM, it didn't carry the actually value information. i.e, no way to get the value as it is already lost in the response.

Daniel · Accepted Answer · 2015-01-06 10:30:29Z

1

If you get the string '[I@48a6ea00' as response, then the one, who has written this int[] datatype into the database, has done something wrong. So it's too late for python or cassandra, to rescue the original numbers.

answered Jan 6, 2015 at 10:30

Daniel

42.9k4 gold badges57 silver badges82 bronze badges

2 Comments

zhihuifan Over a year ago

do you mean it is impossible to convert string into Python List or it is impossible to convert THIS string '[I@48a6ea00' to python List?

zhihuifan Over a year ago

no one write this data to database. I got this string via MX4J (cassandra used this Java library to support some remote call to know the state for cassandra running)

Collectives™ on Stack Overflow

decode Java int array in python code -- found a workaround for my case

2 Answers 2

2 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related