Sometimes PolyAnalyst does not halt its operations immediately after clicking the Stop tasks button. Sometimes it takes several seconds (or even minutes) as PolyAnalyst must perform several cleanup operations. During this time the task list may show no change and PolyAnalyst may appear to continue working as if you had not chosen to stop all tasks. Please be patient.

As an example of why this might take so long, consider the situation where a node is importing several hundred million records, has made some progress, and now has to delete all of that data it imported up to the point of being halted. In addition, in some cases PolyAnalyst may not stop a task even after clicking the Stop tasks button. For example, the task of saving a project cannot be stopped.

Exiting the client while tasks are running does NOT stop tasks from running. Tasks are run by the server and not by the client. When you initiate a task using the client, you are merely sending a message to the server to start a task. The client is just displaying what tasks the server is running. If you close the client while tasks are running and then later reopen the client and reopen the project you were working with, any tasks that are still running in the server will automatically reappear in the task list.

In addition to checking whether the configuration of a node is valid at the time that you click the OK or Execute button in the properties window of the node, or at the time you select the Execute option from a node’s right-click menu, sometimes an incorrect configuration is not determined until the node actually starts to execute. Many nodes perform additional checks to determine whether they can successfully execute around the time they start executing. Therefore, it’s possible for one of these real time checks to fail. Sometimes a node may stop executing and display a large red X over its icon on the flowchart (and possibly in the task list). Right-click on the node and select Display error description. View the error description and determine how to correct the configuration so that the node can eventually execute.

If there is a sign of an error on a node icon, you can open the node log and get information about the type of an error.

A conversion error can occur if PolyAnalyst guesses incorrectly the type of a column while importing the data, for example, in the CSV source node. For example, if a dataset contains 397 integer and 1 numerical value, while importing the node will automatically determine that the column contains data of the integer type (as integer values prevail) with one exception. To correct this error, it’s needed to manually change the type of data to the correct one on the Column specifications tab. To do this, click the Refresh button to display all the columns on the Column specifications tab and change the data type in the User column. To change the data type in a column, double-click on the required cell and open a drop-down menu with all the available data types. Then select the required data type from the list.

A node’s view, if open, is automatically closed when you change the properties of a node or reexecute a node. This can also happen if you modify or execute any node preceding the node in a sequence, because that subsequently causes the node you are viewing to update or execute.

Similarly, if another user modifies the node or executes the node or any earlier node while you have the project open and have concurrent project editing enabled, then the window automatically disappears.

When you decide to view the contents of a node, PolyAnalyst must load that information into your computer’s memory. Most of the time this is an extremely fast process, which leads many users to always expect this to happen quickly, but it is not always the case. When dealing with larger amounts of information, or with a less powerful computer, or with a poor connection to a server that is processing several other things simultaneously and not located on the same machine as the one you are using, you may notice a delay in opening the view. You can view the Sent and Received information in the status bar to view that the program is in fact performing something, even though nothing appears to be happening. If after waiting for a full minute, you still do not see any new window, it may be that you did not click the view, or you may have encountered an error and should contact Megaputer Support.

Should a chart view appear frozen while opening the view, this is typically due to the processing which must occur to render the chart on the screen. When graphing a very large number of classes (more than a few hundred), PolyAnalyst must perform a large amount of calculations and this can take a considerable time. This happens in competition with the process which displays information, which causes PolyAnalyst to appear as frozen. Users need to be patient in this case and wait for PolyAnalyst to complete processing, or consider not trying to graph an attribute with such a large number of distinct values/intervals. It is common practice to avoid attempting to graph such a large number, regardless of the processing constraint, as the output of the graph is difficult to interpret as it is hard to discern among slices and judge the sizes of individual slices.

This is not actually an error, but is instead a common source of confusion. The reason a node cannot be viewed is generally because the node is not in the correct state. Revisit the earlier introduction to node states. Nodes that have not yet completed processing do not produce viewable results, so PolyAnalyst logically prevents you from trying to view such non-existent results. There are several reasons why a node might not be in the completed state. For example, the node may not be configured, or may not have been run. It is always necessary to first configure and then execute a node prior to being able to view it. Even then, if you execute the node, view its results, and then decide to reconfigure the node, its results are deleted at the time you reconfigure the node, so its results will no longer be viewable.

In rare situations involving a large number of columns, or a slower computer, the list of columns may take several seconds to update after typing a new character into the Filter area.

On the average computer, this effect generally only appears when working with a dataset containing more than 500 columns. If you are experiencing a slow column load for a smaller dataset of less than 50 columns, it is a sign that something else, other than the shear number of columns, is slowing down the process.

You can reduce the chances of this happening by filtering out unused columns using a Filter column.