TDBGrid columns design time or run time?

Question

I am using TDBGrid for the first time.

I didn't even notice the Columns property in the object inspector and everything went just fine: the grid was filled.

Then I saw the Columns property and tried it. When I use the object insector to add items to the Columns property of the grid, they do not all of them show up at run time. To me, it looks like the FieldName property is important. It looks like it is correct, but there are no values in some columns.

If I leave this property blank (no columns at all) it seems that the control populates it itself at run time and I do see all the columns. I could live with that. When I set column titles at run time it works, but setting column width does not.

---

[Update] Thanks for the feedback. I am learning slowly. Now I see what is happening. I wanted three of the five columns to be calculated. Two are gotten from INNER JOINs and the third from an SQL 'cast' of SELECT ... FROM_UNIXTIME(<table>.time_stamp).

If I do not define the columns at design time then these values are populated at run time when I execute my SELECT. Cnn anyone tell me how to set things up at design time? Thanks

Answer 1

There is no difference in functionality between runtime and designtime set up columns. That is: what can be done designtime can also be done runtime, and vice versa.

When te Columns property is kept empty, the grid fills it by default with all fields it finds in the linked dataset, once the dataset is opened. All column specific settings, such as alignment, width, title caption, etc... are set to default, based on the type, length and name of the corresponding field. This is the lazy way which sure could be sufficient.

You could deside to use your own setup or modification of the columns property, for one or more of the following reasons:

• Change the order of the dataset fields in which they must appear in the grid,
• Truncate the width of text fields (especially VarChar fields or Char fields with a large length normally result in colums that are too wide for normal presentation and/or user interaction),
• Change the alignment of the fields (normally numeric fields are right aligned by default),
• Hide one or more fields from the dataset,
• Etc...

These own settings can be done desingtime as well as runtime.

Using the designtime columns editor is the preferred and RAD way in need of these customizations. Open the editor by double clicking the grid or the property, or via the grid's context menu. If the linked dataset is active then the command Add all fields is enabled, which does exactly the same as if you did not select any field at all, i.e. the default behaviour as described in the second paragraph.

Answer 2

The best way to define the columns on a dbGrid is to doubleclick the dataset in order to create variables that the form knows about (persistent fields). You can edit various properties of these variables such as DisplayLabel, DisplayWidth and DisplayFormat.

You can also add calculated fields by double clicking on the dialog box which appears when you double click on the client dataset.