Spreadsheet Import

Table of Contents

01. Overview

02. Spreadsheet Import Steps

03. RapidRMA Task Fields

01. Overview

For very simple models, ones that consist of solely tasks running on one or more nodes, one can use the RapidRMA "Spreadsheet Import" feature. The only resources that can be specified are the node CPUs (implicitly specified by specifying a node name for a task). Other resources cannot be specified through this import process.

For more general models, consider using the PERTS Task Graph and Resource Graph files for import (  File | Import | Resource Graph & Task Graph  ). These are text files with a more complex format than the simple one-line-per-task "spreadsheet" format used in the feature we describe here. The graphs allow the specification of all the resources and of all the usage of each resource by each task.

We hope the name "spreadsheet" is not too misleading. There is no interface with any spreadsheet application as far as extracting the data directly from the application's internal format. However, most spreadsheet applications provide a low-level "export" format known as comma- or tab-delimited text files. It is those files that RapidRMA imports through the "Spreadsheet Import" feature.

In these comma- or tab-delimited text files, each line corresponds to a spreadsheet row, and each delimited field in a line corresponds to the a cell in a row. If the fields in all the lines were properly spaced out horizontally, one would see the fields form columns, and each field corresponds to a spreadsheet cell (the intersection of a row and a column). The delimiter character is also known as the field separator character. The delimiter character cannot appear in the content of any of the fields. The first line should have field values that serve as column headers (names that describe the column).

Figure 1 shows an example of a file for import. This file uses commas as field separators. Not all the RapidRMA model fields need to be present (have a column of data in the file). Absent fields are provided with defaults similar to the defaults provided when entering data through the RapidRMA task and resource editors. However, for the columns in this file, every row must have a non-null value for that column. Null values are not allowed in any cell.

Task,Node,Period Type,Period,Period Min,Period Max,Work Type,Work,Comments
MYTASK-10,MYNODE-1,Deterministic,30,30,30,Deterministic,5,First Task
MYTASK-11,MYNODE-1,Uniform,40,50,60,Deterministic,10,Second Task
MYTASK-12,MYNODE-1,Exponential,40,60,60,Deterministic,10,Third Task
MYTASK-20,MYNODE-2,Deterministic,30,30,30,Deterministic,5,Fourth Task
MYTASK-21,MYNODE-2,Deterministic,50,50,50,Deterministic,10,Fifth Task
MYTASK-22,MYNODE-2,Deterministic,60,60,60,Deterministic,10,Sixth Task

Spreadsheet Import Steps

Step 0. Invoke Spreadsheet Import

Start the import of a tab- or comma-delimited file (a format typically exported from a spreadsheet) through the File Menu:  File | Import | Spreadsheet.   Figure 2 shows the path, starting at the menu bar, going down the "File" drop down menu, proceeding through the "Import" cascading menu, and selecting the "Spreadsheet" item in that menu.

Figure 2. Invoking the "Spreadsheet Import" feature through the "File" menu.

In response to this selection, an "Open File" dialog box appears and one selects the exported spreadsheet file to import. There are no preferred file extensions used by "Open File" when scanning a directory for files to offer.

On selecting a file, the import process reads the file and requests information from the user in order to interpret the file. This is accomplished through a sequence of dialog boxes:

• Specify the delimiter

• Match columns to RapidRMA fields

• Final review

Step 1. Select Delimiter

The first import dialog box provides the means for one to specify the field separator character. An example of this dialog box is shown in figure 3 below.

Figure 3. The Step 1 dialog box of the "Spreadsheet Import" feature allows one to specify the field separator (delimiter) used in the file.

The first section at the top is called "Delimiters". It is a list of common field separators (semicolon, comma, and tab) that can be checked, or one can specify a character.

Below the "Delimiters" section is a "Data Preview" section. This section provides visual feedback regarding the choice of the field separator. The "Data Preview" section shows each line of the file as a row. The fields in the lines are aligned in columns. The splitting of the line into fields is based on the field separator chosen in the "Delimiters" section. As one chooses a field separator, the columns in the "Data Preview" section re-aligned themselves.

After specifying the appropriate delimiter, click on the "Next" button to proceed. The navigation buttons at the bottom of the dialog box are:

• Cancel. Abort the import operation and return to the main application window.

• Back. Not available ("grayed out").

• Next. Proceed on to the next step match data columns to RapidRMA model fields.

Step 2. Match Spreadsheet Columns to Fields

The second import dialog box allows one to assign columns to the RapidRMA model fields. An example is shown in Figure 4.

Figure 4. The Step 2 dialog box of the "Spreadsheet Import" feature allows one to specify the mapping of file columns to task fields in the RapidRMA model.

This dialog box consists of one main section that shows three columns. The leftmost column ("RapidRMA Cols") lists the RapidRMA model fields that can be loaded through this feature. All these fields are task attributes. The middle column ("Col Type") shows the datatype of the field. The datatype of most fields is either "String" or "Integer". The datatype of "One of" is discussed below. The right column ("Imported Type") shows the column labels taken from the first line of the import file.

In this dialog box, one assigns the file columns to the RapidRMA fields by dragging a file label cell up or down to place it in the same row as the RapidRMA model field to which it corresponds. The contents of the target and source cells in the "Imported Cols" column are swapped. Also, for extra fields in the "Imported Type" column that do not correspond to RapidRMA model fields, one can delete the label by right-clicking on the cell to pop up a "Remove Imported Column" button, and then clicking on that remove button. The button can be dismissed by clicking elsewhere on the dialog box.

Figure 5 shows the dialog box after dragging the file field "Comments" from the "Amount of Work Min" row down to the "Comments" row.

Figure 5. In a second view of the Step 2 dialog box of the "Spreadsheet Import" feature, the file column "Comments" has been dragged to a different row so that now it corresponds to the RapidRMA task field "Comments".

Three fields ("Period Selection", "Amount Of Work Selection", and "Deadline Type") consist of one selection from a list of keywords. These datatype of these fields is "One Of". Click on a "One Of" cell to display a list of the valid keywords for that cell. This is shown in Figure 6.

Figure 6. The "One Of" datatypes in the "Col Type" column can be clicked to reveal the permitted keywords.

After specifying the appropriate delimiter, click on the "Next" button to proceed. The navigation buttons at the bottom of the dialog box are:

• Cancel. Abort the import operation and return to the main application window.

• Back. Return to the previous step: select delimiter (field separator) character.

• Next. Proceed on to the next step: final review.

Step 3. Final Check

The third import dialog box is shown in Figure 7.

Figure 7. The Step 3 dialog box of the "Spreadsheet Import" feature allows for a final review of the data field mapping before creating the RapidRMA model.

If the import process detects type errors or null values in any fields, the cells are flagged with a yellow background and a message for each yellow cell is shown in the box below the data display. In case of such errors, the "Finish" button will be unavailable and the import cannot be completed. One must go back to the spreadsheet or file, correct the problem there, and start the import process from the beginning.

After verifying the file data properly interpreted and mapped correctly, click on the "Next" button to proceed. The navigation buttons at the bottom of the dialog box are:

• Cancel. Abort the import operation and return to the main application window.

• Back. Return to the previous step: match data columns to RapidRMA model fields.

• Finish. Process the data into the RapidRMA model and return to the main application window with the resulting model loaded.

Recall that the model must be saved (  <File | Save As ...  ) before one can proceed to the Analyzer and run an analysis.

RapidRMA Task Fields

When using the RapidRMA "Spreadsheet Import", values must appear for every field in the comma- or tab-delimited file. The import process will not accept null (empty) values.

The current "Spreadsheet Import" feature is our first release of this feature. We need to improve the error checking to check and edit the values to avoid inconsistencies. Through the "Spreadsheet Import" feature, one can set values in a model that could not be introduced when specifying values with the task editor. Later, if those values are editted with the task editor, the values will not be acceptable and the task editor will complain. Worse, some inconsistencies that could not be introduced through the task editor are not checked for at all, and it is conceivable that unidentified inconsistencies could result in an incorrect analysis.

In the list of RapidRMA fields below, we discuss special considerations for values in order to maintain the model in an internally consistent state. For the general meaning of these fields in the model, see the regular RapidRMA documentation.

Task Name (String)

Active Resource (String)

The names in this column are the names of nodes where the task runs. The task is assigned to that node's "CPU" resource.

Period Selection (One of "Deterministic", "Uniform", "Exponential")
Period (Integer)
Period Min (Integer)
Period Max (Integer)

The "Period Selection" determines how the "Period", "Period Min", and "Period Max" values are interpreted and used in the model. There are some peculiarities about how the values are treated, so treat these with some care to avoid inconsistencies. Normally these values would be obtained through a dialog box that ensures the fields are loaded consistently.

For example, with deterministic distribution, one numeric field (main without the max or min) are offered. Internally the max and min are loaded with the main value. With the uniform distribution the max and min (without the main) are offered. Internally the main is set to the mean of the max and min. Through the spreadsheet on can specify a deterministic distribution with different values for the main, min, and max. This inconsistency may show up through some error message later and, in any case, it might be the source of a bug in the user's model. So this "Spreadsheet Import" feature should be used with caution.

If all the tasks are deterministic in period, then it is probably best to have the main "Period" field alone and not have the "Period Min" and "Period Max" fields at all. However, if the tasks have an assortment of period types so that the two or more of the period fields are required, then we suggest you try to keep them consistent in following way.

• Deterministic Periods: Make the "Period Min" and "Period Max" the same as the "Period".

• Uniform Periods: Make the "Period" the average of the "Period Min" and "Period Max".

• Exponential Periods: Make the "Period" the same as the "Period Max".

• Be sure that, for each task, the "Period Min" is less than or equal to the "Period" and the "Period" is less than or equal to the "Period Max".

Amount Of Work Selection (One of "Deterministic", "Uniform", "Exponential")
Amount Of Work (Integer)
Amount Of Work Min (Integer)
Amount Of Work Max (Integer)

See the comments for the "Period" fields.

Ready Time (Integer)

Relative Deadline (Integer)

The "Relative Deadline" should be less than or equal to the "Period".

Drop Deadline (Integer)

The "Drop Deadline" should be greater than or equal to the "Relative Deadline".

Phase (Integer)

Priority (Integer)

Nominally this field specifies user-defined priorities. The value "-1" means the priority should be assigned by the RapidRMA model analysis.

Network QoS (Integer)

Deadline Type (One of "Hard", "Hard Deadline" "Soft", "Soft Deadline" "No Deadline")

Instances (Integer)

Comments (String)

The minimum set of attributes required for an import file are "Task Name", "Period", and "Amount of Work". The distribution for period and work default to "Deterministic" ("Period Selection", "Amount of Work Selection").