VI Builder Output Object

The Visual Integrator (VI) Builder output object creates a DI Model from the data flow.

The Builder output object has four panes where you set attributes.

VI Builder Object All Panes

Object Attributes

Set attributes for the Builder output object in the object attributes pane.

The attributes are organized in the following categories:

Common Options

Example of the common options for a VI Builder output object

Attribute	Description
Input	Defines the object from which the data flow arrives. Use one of the following methods to attach the Builder object to an input data flow object: Right-click the data flow object, and click New Object after this > Output > Builder. Place the Builder object on the task flow, and click its Input attribute's browse button. Move the pointer to the canvas and note that a dotted-line connector is attached to the Builder object. Attach this connector to the data flow object. NOTE: The Builder object's Input attribute is populated with the name of the input data flow object.
Output	Defines the name and path of the model that this object creates. Enter a path and file name in the Output box, or use the browse button to select the file. If a Description attribute is defined, this attribute is not necessary.
Journal	Defines the name and path of the journal file that records parse errors in the input and provides a record of the dump key information. Enter a path and file name in the Journal box, or use the browse button to select the file. This file is displayed in the Task List under LogFile. If not indicated, a file named build.jou is placed in the same directory as the VI script.
Builder	Points to the location of builder.exe. Workbench should automatically find the executable file. If necessary, click in the Builder box and select the executable file if one is listed, or click the browse link and navigate to the executable file.
Description	Specify the name of a Build Description file. Enter a path and file name in the Description box or use the browse button to select the file. A Build Description file is a configuration file that controls the actions of the build. Using the Build Description file provides a way to easily repeat builds without having to retype each option. If a Description attribute is defined, the Output attribute is not necessary.
Defines	Defines values for existing parameters that are defined in the Build Description file. This attribute is only used if you are using a Description attribute and the specified Build Description file has parameters defined. The Defines attribute displays the number of parameters defined. It defaults to [Show (0) defines]. Click the browse button in the Defines box to open the Defines for <object_name> dialog box. Enter the parameters in the following format: `<parm-name>=<parm-value>`

Model Structure

Example of model structure attributes for a VI Builder output object

By default, summary types are not calculated and only the total summary is calculated.

Attribute	Description
Sum_Type_Maximum	Defines whether to calculate the maximum number of summary types. True—Maximum number of summary types are calculated. False (default)—Maximum number of summary types are not calculated.
Sum_Type_Minimum	Defines whether to calculate the minimum number of summary types. True—Minimum number of summary types are calculated. False (default)—Minimum number of summary types are not calculated.
Sum_Type_Std_Dev	Defines whether to calculate the sum of the squares of the column values.. True—Sum of the squares of the column values is calculated. False (default)—Sum of the squares of the column values is not calculated.
Index_Dimension	Creates a single-dimension model that is suited for lookups or single-dimension index models. No other dimension selections in the Model Structure are allowed when using this attribute.
Index_Model	Defines whether to create an index model. true—Creates an index model, which is the same as defining every column in the Detail Model Structure as detail dimensions. Because this attribute defines a Detail Model Structure, it conflicts with setting detail dimensions in the Detail Model Structure tab in the Column Grid. false (default)—Does not create an index model.

File Options

Example of the File Options attributes for a VI Builder output object

Attribute	Description
Journal_Append	Defines whether to append build information in the journal file or to overwrite it. true—Build information is appended to the existing journal file. false (default)—Build information overwrites the previous journal file.
Bucket_Dir	Specifies the directory used to store the temporary sort files. By default, buckets are stored in the current directory. Click the browse button to select a different directory.
Output_Dict	Creates a dictionary file that describes the file created by the Output_Text attribute. Enter a path and file name in the Output_Dict box, or use the browse button to select the file.
Output_Text	Creates a new file containing the input data from the current input file with just the columns used in the build. This attribute is useful with calculated fields, as the text file output contains the calculated field values for each line. Enter a path and file name in the Output_Text box, or use the browse button to select the file.

Tuning Options

Example of tuning options for a VI Builder output object

Attribute	Description
Blocking_Factor	Defines the amount of pre-processing. Pre-processing allows ProDiver to respond faster. The larger the block factoring, the less pre-processing is done. The default value is 100000 rows. The list has the following values: 100000, 50000, 20000, 5000, and 10%. You can enter another value if required. If the blocking factor is expressed as a percentage, the block factor is re-calculated for the model being built, and then the Blocking_Factor_Min and Blocking_Factor_Max attributes become available.
Blocking_Factor_Min	Defines the minimum blocking factor to use when the Blocking_Factor is given as a percentage of the input records. If the calculated block factor is less than this minimum, the minimum is used instead. The default minimum is 10,000.
Blocking_Factor_Max	Defines the maximum blocking factor to use when the Blocking_Factor is given as a percentage of the input records. If the calculated block factor is greater than this maximum, the maximum is used instead. There is no default maximum.
All_Skip_Values	Specifies dimension values to skip when summarizing data. When Builder summarizes data for the specified dimension value, it uses the large blocking factor, optionally defined by the Skip_Blocking_Factor attribute. This reduces the amount of pre-processing done for these data values, but diving into these values is slower. Use this option for dimension values that you do not expect users to dive on. To define values, click the browse button to open the Skip values for <object_name> dialog box, and enter dimension name and value in the following form: `dimension:dimension-value` Just the dimension names appear in the All_Skip_Values attribute, not the values. If you set the dimension to an asterisk (*), the specified value is used as skip value for all dimensions.
Skip_Blocking_Factor	Specifies the block factor used for skipped dimension values. This attribute defaults to the number of records, which is the largest possible value.
Working_Space	Specifies the amount of available space for swapping information during the build. The default value is 30000000 (30 MB). Unless the file is large enough to cause errors, DI recommends not changing this value.

Control Options

Example of Control Options attributes for a VI Builder output object

Attribute	Description
Model_Encoding	Defines the encoding used to create the model: auto—The model is encoded based on the Unicode state of other objects in the same task. If no other object uses Unicode data and the script itself is not encoded as Unicode, the data is stored in the model as ISO-8859-1 or Latin1 characters. ascii—The model is encoded as ISO-8859-1 or Latin1 characters. latin1 (default)—The model is encoded as ISO-8859-1 or Latin1 characters. utf-8—The model is encoded as UTF-8 Unicode characters. unicode—The model is encoded as UTF-8 Unicode characters.
Dump_Key	Assists in debugging. true—A profile is generated to a journal file. false (default)—No profile is generated.
Keep_Buckets	Indicates if temporary sort files, referred to as bucket files, are kept. true—Temporary bucket files are kept. false (default)—Temporary bucket files are deleted.
Parse_Only	Determines if the Builder object checks the input file for errors without creating a model. true—The Builder object checks for errors. false (default)—The Builder does not check for errors.
Quiet	Indicates whether to suppress the display of phase progress information during the build. true—The phase progress information is suppressed. false (default)—The phase progress information is displayed in the log.
Remove_On_Fail	Specifies the action for a failed build. true—If the model build fails, the model is automatically deleted. This prevents users from accessing incomplete or imperfect data sets. The user receives a Model did not finish message if they attempt to open a failed model. CAUTION: When building a model, any existing model of the same name is overwritten by the Builder object with no records, and the build work continues. Failure to finish successfully does not restore the original model. On Windows, if the model is open in DiveLine when a rebuild fails with this option, it does not disappear from disk until DiveLine closes it. false (default)—If the model build fails, the model is not deleted.
Squash	Specifies how to handle duplicate rows. true—Duplicate rows are collapsed while totaling the summaries. This reduces build time, memory requirements, and the space required for the model when there are many duplicate records. Using this option deactivates the calculation of maximum, minimum, and standard deviation columns in the model. The count of records in the model reflects the number of records after the squash. false (default)—Duplicate rows are not collapsed.
No_Build	Specifies whether the Builder processes options without building a model. true—The Build object processes options but does not build a model. This is useful to create or test a new build description without building a model. false (default)—The Build object produces a model.
Built_In	On Windows servers, determines whether to use the internal or external Builder. true (default)—VI runs Integrator and Builder in a single process using the internal Builder component to build the model. While more efficient, it does not allow for multiprocessing. false—VI uses Integrator to process the script but uses the external Builder to build the model.
Allow_Empty	Defines how to handle scripts that return no records due to the parameters that are used to filter the data. true—Allows empty models. false (default)—Does not allow empty models.

Model Variable Options

Example of Model Variable Options attributes for a VI Builder output object

Attribute	Description
Extract_Time	Describes when the input data was extracted from the source data. If blank, the extract time is set to the build time. If set, use one of the following formats: YYYY-MM-DD HH:MM:SS YYYY/MM/DD HH:MM:SS If no time is specified, the time is set to noon of the date. For more information, see the ProDiver Help > About Text Variables. NOTE: The build time is recorded automatically and is available in Diver Reports using $BUILD_DATE and $BUILD_TIME.
Comments	Specifies a comment that is available from the ProDiver Open File dialog box's Get Info button. Limit is 100 characters. This can also be accessed in ProDiver using the report variable $COMMENTS
Footer	Specifies a string to appear at the bottom of every page printed from ProDiver. It is accessed by using the variable $MODEL_FOOTER in the ProDiver report or report palette.
Model_Vars	Specifies the user-defined model variables for the model. Click the browse button to open the Model_vars for <object_name> dialog box. Variables use the following form: `<variable_name>=<variable_value>` These model variables are available in ProDiver through the $MODELVAR reporter variable.

Incremental Build Options

Example of Incremental Build Options attributes for VI Builder output object

Attribute	Description
Update	Indicates whether to update the model with incremental build data, which places additional data in a separate file with an mdu extension. When the model is opened, the mdu data is seamlessly included in the model. Repeated use of the Update attribute increases the size of the mdu file. true—Updates use an incremental build. false (default)—Updates do not use an incremental build.
Compact	Indicates whether to do a full build that merges incremental files with the main model and recovers space from incremental builds. true—Performs a full build of the model to recover lost space from incremental builds. false (default)—Does not perform a full build.

Attribute

Description

Update

Indicates whether to update the model with incremental build data, which places additional data in a separate file with an mdu extension. When the model is opened, the mdu data is seamlessly included in the model.

Repeated use of the Update attribute increases the size of the mdu file.

true—Updates use an incremental build.
false (default)—Updates do not use an incremental build.

Compact

Indicates whether to do a full build that merges incremental files with the main model and recovers space from incremental builds.

true—Performs a full build of the model to recover lost space from incremental builds.
false (default)—Does not perform a full build.

Input Format Options

Example of Input Format Options attributes for a VI Builder output object

Attribute	Description
Ignore_Line_End	Specifies whether or not parse errors dealing with the end of an input line are ignored. For example, Fixed field occurs past end of line or Too few fields. This attribute helps control processing when there are too few data items in an input row. true—Processes the line false (default)—Prints out a warning message, and skips the line
Ignore_Parse_Errors	Specifies whether to ignore parse errors during the build. true—Parse errors encountered in the input file, such as extra fields and string values in numeric fields, are flagged, but the build continues. The lines with parse errors are ignored. This attribute is useful for running preliminary tests on data and allows builds to continue while the data is cleaned up. false (default)—Parse errors are not ignored.

Attribute

Description

Ignore_Line_End

Specifies whether or not parse errors dealing with the end of an input line are ignored. For example, Fixed field occurs past end of line or Too few fields. This attribute helps control processing when there are too few data items in an input row.

true—Processes the line
false (default)—Prints out a warning message, and skips the line

Ignore_Parse_Errors

Specifies whether to ignore parse errors during the build.

true—Parse errors encountered in the input file, such as extra fields and string values in numeric fields, are flagged, but the build continues. The lines with parse errors are ignored.

This attribute is useful for running preliminary tests on data and allows builds to continue while the data is cleaned up.
false (default)—Parse errors are not ignored.

Column Grid

You set attributes for the output model in the Model Structure column grid.

For each input column you can indicate the following column properties.

Attribute	Description
Input Column	Displays the name of each input column. This attribute is read-only.
D (Dimension)	Defines the input column to be a dimension in the model output. At least one input column must be defined as a dimension. To define a column as a dimension, click the D column for that row.
S (Summary)	Defines the input column to be a summary column in the model output. To define a column as a summary column, click the S column for that row.
I (Info Field)	Defines the input column to be an info column to a corresponding dimension in the model output. To define a column as an Info Field, click the I column for that row.
DC (Dimension Count)	Instructs the build to precalculate the dimension counts (DimCounts) for a dimension. For more information, see Column Properties Dialog Box.
Date	Defines the input column as a date. It is important to identify date columns so that they are stored properly in the model, facilitating ProDiver functions such as sorting, date and time functions, and timeseries. To define a column as a date, click the Date column for that row. For more information, see Column Properties Dialog Box.
Props (Column Properties)	Opens a separate Column Properties dialog box where you can set detailed properties. For more information, see Column Properties Dialog Box.

Column Properties Dialog Box

The Column Properties dialog box allows you to set detailed properties for dimension, summary, and info field columns prior to building the model.

Example of Column Properties dialog box for a VI Builder output object

Attribute	Description
Name	Displays the name of the input column
Alias	Sets an alias for the individual column. Select the check box and type an alias for the column. Spaces before or after this name are ignored, but spaces within the name are allowed.
Column is a string	Columns with numeric data are considered numeric columns and any leading zeros are stripped off. Select this check box for columns that contain numeric data that should not have any calculations made on it, such as zip codes or part numbers. This attribute preserves leading zeros in the data and allows use of string manipulation functions.
Column is a date	Select this check box to treat the column as a date. This attribute is equivalent to clicking the Date column for an input column in the Model Structure. If selected in the Model Structure, the Column is a date check box is selected and the same in reverse. When creating a model, Builder converts date columns into an internal date format (an integer number). This number represents the count of days since December 31, 1899. This internal format lets ProDiver easily manipulate dates for use in time series, period comparisons, calendar plots, and dimension date roll-up features. This allows ProDiver to sort dates chronologically, by year as well as month and day. For example, 12/10/2011 would come before 10/10/2012, but after 01/01/2011.
Format	If the Column is a date attribute is selected, choose a date format from the list. Any date prior to January 1,1900 displays a Date Out of Range message. Choices include: `YYYY/MM/DD` `YYYYMMDD` `MMDDYYYY` `MM/DD/YYYY` `YYYY-MM-DD`
Period calendar	N/A
Date start year	If the Column is a date attribute is selected, use this attribute to specify the beginning date range for handling two-digit years within a 100-year window. This attribute allows the Builder object to correctly interpret two-digit years within the specified range. The Builder object then converts the two-digit years to the appropriate four-digit dates. For example, specifying 1950 indicates a range between 1950 and 2049 such that 79 gets interpreted as 1979 and 14 gets interpreted as 2014.
Column contains nullable data	For columns marked as summaries, select this attribute so that Builder distinguishes between zeros and nulls and keeps a count of the non-null values in the input. Normally, Builder treats null values in input as zeros. This count is used when calculating averages and standard deviations on the field in ProDiver. Setting this attribute slightly increases the size of the resulting model.
Column Precalculates DimCounts	For columns marked as dimensions, select this attribute for Builder to precalculate dimension counts (DimCounts). This attribute is equivalent to selecting the DC column for an input column in the Model Structure. If DC is selected in the Model Structure, the Column Precalculates DimCount check box is checked and the same in reverse. In ProDiver, DimCounts display the number of unique values for all dimensions, based on the selected dimension value. It is a numeric value that displays in the ProDiver console, but it can also be added as a calculated column in a tabular. DimCounts deliver a large amount of information with very little work. However, they are computation-intensive, and can be slow for large data sets with many unique values in the dimension counted. Having Builder precalculate some DimCounts can speed up ProDiver access to DimCounts for non-grouped core dimensions.
Sort by	For columns marked as dimensions, allows you to set up an alternative sort. By default, ProDiver sorts low to high in the following order: Numbers Uppercase characters Lowercase characters This alternate sort is an internal sort within the model of one column following the default sort of another column. An example of an alternate sort is to sort the dimension parts by part number rather than alphabetically. Select the desired column from the list to change the sort order for the dimension.
Skip during the build	Specify one or more dimension values to skip when summarizing data. When Builder summarizes the data for the dimension value, it uses a large blocking factor. You can optionally specify the Skip_Blocking_factor attribute to reduce the amount of preprocessing Builder does for these data values. However, diving into these values is slower. You should use this option only for dimension values that you do not expect users to dive on. To set the dimensions and values to skip In the Skip during the Build box, enter the dimension name and value in the form: `<dimension_name>:<dimension_value>` Click the plus sign (+) button. The dimension and value are added to the box. Repeat for any additional dimension values. To remove a dimension, select it in the box and click the minus sign(-) button.