"Read Constructions Operator filtering problem"

Cleo

Hello

I have created a very simple rapidminer process which loads data from a database then calls the Write Constructions operator. 

I then have a second rapidminer process which loads the exact same data, which is passed through a Read Constructions operator (which loads the file previously created) and then to an excel file.  This excel file is blank.  The Read Constructions operator seems to be filtering out all the attributes.
I have got this process to work before, but for some reason I am having a lot of difficulties getting this to work now.  Also I am using RapidMiner 5.0.1

Any suggestions
Thanks
Cleo

PS here is the contents of the Write Constructions Operator

<?xml version="1.0" encoding="windows-1252"?>
<constructions version="5.0">
    <attribute name="timestamp" construction="timestamp"/>
    <attribute name="date" construction="date"/>
    <attribute name="open" construction="open"/>
</constructions>

land

Hi Cleo,
these constructions doesn't seem to do anything? In fact they are constructing three attributes as itself. This is not too sensible. Is that really what they should do?
Normally Constructions should contain formulars for calculating new attributes from combinations or transformations of old, but this need some special operators, which will perform this during training. As long as you simply load the data, there will not be any construction beside saying: "Hey, I'm myself"

Greetings,
  Sebastian

Cleo

Hello Sebastian,

I do realize this is a simple example, however after running the data through a Read Constructions the original attributes seem to disappear.

Jan 26, 2010 10:51:14 PM WARNING: Kernel Model: The number of regular attributes of the given example set does not fit the number of attributes of the training example set, training: 5, application: 0
Jan 26, 2010 10:51:14 PM WARNING: Kernel Model: The given example set does not contain a regular attribute with name 'open'. This might cause problems for some models depending on this particular attribute.
Jan 26, 2010 10:51:14 PM WARNING: Kernel Model: The given example set does not contain a regular attribute with name 'macd'. This might cause problems for some models depending on this particular attribute.
Jan 26, 2010 10:51:14 PM WARNING: Kernel Model: The given example set does not contain a regular attribute with name 'momentum'. This might cause problems for some models depending on this particular attribute.
Jan 26, 2010 10:51:14 PM WARNING: Kernel Model: The given example set does not contain a regular attribute with name 'ma'. This might cause problems for some models depending on this particular attribute.
Jan 26, 2010 10:51:14 PM WARNING: Kernel Model: The given example set does not contain a regular attribute with name 'attribute_13'. This might cause problems for some models depending on this particular attribute.

<?xml version="1.0" encoding="windows-1252"?>
<constructions version="5.0">
    <attribute name="open" construction="open"/>
    <attribute name="macd" construction="macd"/>
    <attribute name="momentum" construction="momentum"/>
    <attribute name="attribute_13" construction="attribute_13"/>
    <attribute name="ma" construction="ma"/>
</constructions>

If I keep all the original attributes I get the warning

Jan 27, 2010 8:55:38 AM WARNING: Kernel Model: The number of regular attributes of the given example set does not fit the number of attributes of the training example set, training: 5, application: 13

Cheers,
Cleo

land

Hi Cleo,
you are right. Exactly this test case does not work. This stupid operator removes all old attributes, even if they are part of the construction. I will correct this behavior.

Greetings,
  Sebastian

Cleo

Thanks,
Cleo