The BANKS system helps you to browse and search data stored in relational
databases conveniently via the web. Other kinds of data such as interlinked
documents or XML can be preprocessed into a form that BANKS can operate on.
You can extract information in a simple
manner without writing complex queries but by typing a few keywords, following
hyperlinks and interacting with controls on the displayed results.
It also allows you to define views that you find interesting with a facility
to store it for later use.
Features of BANKS
Browsing
BANKS provides a rich interface to browse data stored in the relational
database. The interface shows the list of available tables in the database and
the auxillary tables.
Selecting a table link allows you to view the schema of the table. You can
either view all records currently stored in the table via "Get all records"
link or a "Select" interface is provided to select records that you wish to
retreive by specifying appropriate values for fields in the table.
Each table displayed comes with a variety of tools for
interacting with the data, namely:
facility to drop a column from the view. data can be sorted on the specified column in ascending or descending
order. data can be restricted to tuples that satisfy certain selection condition
on the selected column. clicking on "join" results in the display of join between the referenced
table and the displayed table on the foreign key column. data can be nested, nesting on a column results in only the distinct values
for that column being displayed. You can click on any of the values to see
the tuples associated with that value. For every attribute that is a foreign key, a link is created in the display
to the referenced tuple. In addition, primary key columns can be browsed
backwards, to find referencing tuples, organized by referencing relation (a
specific referencing relation can be selected by you). next and previous buttons help with browsing.
Templates
Templates provide several predefined different ways of displaying
the data. You need to customize the template by specifying some information(
depending on the template) and then it is made available for you to
browse.
Currently BANKS provides four types of templates:
Crosstab
Crosstab allows grouping of records on two specified dimensions;
values for one dimension for the rows and values for the other form the
columns of a crosstab, as can be expected. The value in each cell is a
summary (count by default), and in addition clicking on the summary in a cell
shows the detailed records.
Nested
This provides for multi-level nesting of data, by specifying a
sequence of nesting attributes. For example, nesting by the department and
program on a student relation allows you to click on the desired
department, after which available programs in the department are shown, and
clicking on a program shows all students in that program in the selected
department.
Folder
This is similar to nesting, but shows the values of the next
attribute in place, instead of in a new screen. Quite similar to folder view
of files and directories in windows explorer.
Graphical
This displays information in a bar chart, line chart or pie chart
format. Links are provided on the image so that clicking on say line of the
line chart or bar of the bar chart, shows the tuples with the associated
value.
Search
BANKS provides a keyword based search facility. You can type in one or more
keywords that relate to the tuples that you wish to retreive. The search will
return tuples containing each of the keywords( in any column), and will output
those tuples whose nodes are in close proximity to the specified keywords.
Result rows are given a relevance based on proximity, and rows with higher
relevance are displayed first. The output displays each resultant tuple and
below it displays its relevance score.
BANKS search facilty also provides for techniques to prune the search.
To summarize BANKS is a user friendly integrated browsing and keyword
querying system. It easily allows users with not much knowledge of databases
to query and browse data.