In this article
You can have more than one item in both the rows and the columns of a table. You can for example add another single question, in this case q7 (Favorite), to the Rows dimension .
Note: You can also include items from more than one data source (project) in the table (go to Override Other Projects for more information).
If you place the second question below the first, the items will be stacked.The first part of the table will then show the results for the Gender question, and below that, the results for the Favorite question. Both of the sets of results will be crossed with the content of the Columns dimension,in this case the Age question.
Figure 1 - Dragging the Favorite question into rows
Click Refresh to update the table and see the result .
Figure 2 - Stacking elements in rows
Similarly, you can stack elements in columns by placing them next to each other. For example, move the Gender question from rows onto columns next to the Age question .
Figure 3 - Dragging Gender into columns
Then click Refresh to see the result .
Figure 4 - Stacking elements in columns
Items can also be nested both in columns and rows. You can, for example, move the Age question above Gender . Again, there is a gray-shaded area that will help you determine where the item will be inserted.
Figure 5 - Dragging Age on top of Gender
Click Refresh to update the table. Note that each Age column is now divided into the genders, and each has its own “Totals” column with an additional “Totals” column for the entire table on the right . Note that for this table the Vertical Percents distribution has been deselected to reduce the size of the table (go to Distributions for more information).
Figure 6 - Nesting elements in columns – Age on top
If you place Gender on top of Age, the table layout changes again. This time the Gender columns are divided into the various ages .
Figure 7 - Nesting elements in columns – Gender on top
A similar nesting with Gender and Favorite in columns will give the following result .
Figure 8 - Nesting elements in rows
Nesting and stacking in Reportal is very flexible, and you can stack and nest variables in several levels. For example, with Age and Gender stacked in columns, you could nest a third question on top of either of them or on top of both. As you move the object above the other items in columns, different areas will be shaded to show where it will be inserted.
Dropping the Favorite question above Age and Gender while the entire width of the area is shaded, as shown in the figure below-left, will give Favorite nested above both Age and Gender .as shown in the figure below-right.
|
Figure 9 - Dragging Favorite above Age and Gender |
Figure 10 - Favorite nested above Age and Gender |
Dropping Favorite above Age while only the smaller area is shaded, as shown in the figure below-left, will nest Favorite above Age, and leave both stacked beside Gender .as shown in the figure below-right.
|
Figure 11 - Dragging favorite on top of age |
Figure 12 - Favorite nested above Age |
It is however, not recommended to nest and stack too many variables in too many levels, at least not if they have long answer lists, as the size of the tables produced will quickly become very large. If the report is run directly against the survey database instead of using BitStream files (go to BitStream Files Overview for more information), the database queries needed to compute the content of the table may then timeout (go to Query Timeout for more information). It is also difficult to read tables that contain too many variables.
Note: If your report is based on BitStream files, you can include the same question in both the rows and columns of the table. If your report does not use BitStream files, this will give an error message.
Note: You can cross a Formula object with a Statistics object, but you cannot cross Statistics with Statistics; if you attempt to do this an error message will be displayed.
You may have noticed in the examples in this area of the documentation that there are no answers for the "Under 18" group in the favorite question (0 counts). This is due to screening: At the beginning of the survey respondents under 18 were prevented from going further, and consequently were not asked the favorite question. In most circumstances, as this column provides no useful information, you would probably wish to remove that column from the table. Do this using masking (go to Answer and Scale Masks for more information) or by removing columns with zero responses (go to Remove Empty Headers for more information).