HBase Schema Design for Parametric Search

[Bin Wang <bi...@gmail.com>]

I have asked this question on Stackoverflow
<http://stackoverflow.com/questions/30879713/hbase-schema-design-for-parametric-search>,
but I have not got a satisfying answer yet. I am trying my luck on the
mailing list hoping someone can help me on this.

I have a sales table contains invoice date, customer information and all
the parametric data for the product sold, like RAM size, DISK, ... for a
laptop.

I am trying to design a schema so I can transform the data into HBase and
build a search tool on top of it.

(1). Can anyone draw a skeleton of what is the proper schema for this type
of search applications other than the naive format like this:

    rowkey: transaction_id
    columnfamily1: transaction_parameter
        column_qualifier1: customer_name
        column_qualifier2: selling_price
        column_qualifier3: selling_cost
    columnfamily2: production_parameter
        column_qualifier1: product_model
        column_qualifier2: disk_size
        column_qualifier3: ram_size
        ...

(2). Is it realistic to do something in HBase to get all the distinct
parameters to populate the parameter dropdown list in the front end.

I am really excited to learn HBase and any guidance or study material will
be much appreciated!

Best regards, Bin