Blog http://forums.visokio.com/categories/blog/p5/feed.rss Mon, 30 Oct 17 12:12:01 -0400 Blog en-CA New formula functions (2.7 b105+) http://forums.visokio.com/discussion/1085/new-formula-functions-2.7-b105- Thu, 03 Nov 2011 15:16:43 -0400 steve 1085@/discussions DATE(year, month, day, hour, minute, second, millisecond, timezone)
For creating date values using numerical values.
E.g. DATE(2007, 2, 1) for 1st Feb 2007.

ROWINDEXSUBSET(rowIndex, operator)
For specifying subsets by row index.
E.g. ROWINDEXSUBSET(5) for the 5th record only
E.g. ROWINDEXSUBSET(CURRENTROW, "<=") for all records up to and including the current record.
These can be combined using UNION and INTERSECTION as needed, before being supplied to one of the subset-accepting functions such as SUBSET_MEAN.

Both are scriptable.]]> Database bookmark filtering (2.7) http://forums.visokio.com/discussion/1070/database-bookmark-filtering-2.7 Tue, 25 Oct 2011 05:48:09 -0400 chris 1070@/discussions
You may or may not be aware that you can create a bookmark to a database in DataManager. After bookmarking you can expand the bookmark to see all of the tables inside that database, organised by catalog and schema. You can then choose which tables you want to work with by dragging them onto your DataManager workflow.

We have now introduced the ability to filter which tables are available. This will be useful if you want to bookmark a database with a large number of tables but only show a subset of these tables.

To use this functionality, create a new database bookmark by selecting "add > Database" in the "Bookmarks" tab or edit an existing bookmark by right-clicking and choosing "Configure". Now click on the "Filters" tab. You can create any number of filters and can filter on the Catalog, Schema or table name. For example if I only want to retrieve tables that are prefixed by "tbl" I can add the filter "Table starts with tbl".]]> Salesforce connector (2.7) http://forums.visokio.com/discussion/1036/salesforce-connector-2.7 Tue, 04 Oct 2011 12:38:05 -0400 chris 1036@/discussions
We have now upgraded the connector to use the new "Bulk Salesforce library" from Salesforce. This should, in theory, significantly improve download speeds. So far we have inserted around half-a-million records into a test table, and have been able to download this data into Omniscope in around 30 minutes. We would be interested to know if you are able to download similar volumes of data and whether you encounter any problems during the download process.

The connector interface is very similar to the Database connector interface; you can select a table or enter a custom statement (Salesforce Object Query Language) and shortly you will have the ability to add one or more filters.

The old Salesforce connector is still available, however you need to turn on legacy features to see it.]]> R Statistics (2.7) http://forums.visokio.com/discussion/1027/r-statistics-2.7 Fri, 30 Sep 2011 07:18:09 -0400 steve 1027@/discussions
Feel free to explore it and the "Data > R Statistics > R operations library" dialog.

Behind the scenes, it works be taking one or more input datasets and optional input parameters, exporting the data and values to R as variables, executing the R script, and importing the data back, which becomes the output of the operation/block.

You can either build up a library of scripts which you use later, or can cut your own scripts ad-hoc in each block. We'll be providing (soon) a set of bundled scripts with Omniscope for performing common statistical tasks such as clustering, so you can perform these tasks without any knowledge of R.

When you design a script, you specify how many input datasets it has, and how many input parameters (and their types / configuration). You then write the script, with the corresponding variable names listed at the side.

When you use a script in DataManager, you pick from the library, link up one or more input blocks, and customise the values for any parameters that script has, before clicking Execute. For example, k-means clustering would have 1 input dataset, which must already be normalised, and would require you to tick the fields to perform clustering upon, resulting in the original data table with additional fields containing the cluster number and other statistical results.]]> Normalise (2.7) http://forums.visokio.com/discussion/1028/normalise-2.7 Fri, 30 Sep 2011 07:21:18 -0400 steve 1028@/discussions
It uses the standard score: (x - mean) / stddev
and allows you to choose whether to convert existing fields or add adjacent new fields.
]]> DataManager: New launch parameters (2.8+) http://forums.visokio.com/discussion/1594/datamanager-new-launch-parameters-2.8- Tue, 12 Jun 2012 07:34:19 -0400 steve 1594@/discussions
  • File source: location path (2.8+)
  • URL source: URL address (2.8+)
  • Database block: database/instance/table names and host/port (2.8+); server-side filtering; within custom SQL
  • Access database: file path (2.8+)
  • Field organiser: default value, replace value
  • Record/field filters: match value
  • Input switch: input selection (accepts text, boolean, number)
  • Search/replace: search & replace values
  • List of values: the list itself


This means you can use a variety of mechanisms to dynamically update these form fields:
]]>