InfoQube IM - Community

Submitted by markfoley on 2009/06/04 05:13

Hi guys, I got a bit lost near the end of our chat about alphabetic filters, but am slowly 'getting it'!

One question though: what does "Grid -> Filter criteria applies to sub items" do?

If you have items like so in a grid:

one

|__two (a subitem)

- If you use a column filter button to filter for "two" it shows the 'two' record

- if you use the filter bar next to the source box - (ie like "two") or similar it does not show the 'two' record

This holds true regardless of whether "Filter criteria applies to sub items" is on or off. Where does it have an effect?

Comments

>One question though: what does "Grid -> Filter criteria applies to sub items" do?

"Filter criteria applies to sub items" means exactly that : it will apply to sub items.

So, expanding from your example.

using an example that's slightly more complex (for convenience)

one

|__two (a subitem)

|__three (a subitem)

four

If you write :

item like "one"

in the filter text box, without enabling "Filter criteria applies to sub items", then refresh (don't forget that), you'll get :

one

|__two (a subitem)

|__three (a subitem)

because only items (not subitems) are getting filtered against the "one" string

But if you write :

item like "one"

in the filter text box, while enabling "Filter criteria applies to sub items", then refresh (don't forget that), you'll get :

one

(sub items are getting filtered against the "one" string)

Again, taking from your example. If you write :

item like "two (a subitem)"

in the filter text box, without enabling "Filter criteria applies to sub items", then refresh (don't forget that), you'll get :

NOTHING

because no item corresponds to "two (a subitem)"

But if you write :

item like "two (a subitem)"

in the filter text box, WHILE enabling "Filter criteria applies to sub items", then refresh (don't forget that), you might get :

one

|__two (a subitem)

BUT ONLY IF "two (a subitem)" also meets the source. and this is where it gets a tad complex because one tends to assume that all items/subitems in a grid meet the source, but they don't necessarily. Only the parents (top/first level items) must...

(And this is one of the mechanisms that 1- most users will fail to grasp and will struggle with (why is this item not in the grid anymore ? why is this item in the grid and not meeting the source ? etc.), 2- will make their item search more difficult. And that's why I use a different strategy than most users here as far as how I use one specific source field for each grid, in most of my grids, that is propagated to all items, in addition to another source field that "characterizes" the items that are actually of a certain "grid type"/category -- I think I described my method in another thread. It seems complicated, but it's not, and it makes the whole IQ experience -- especially searching -- more flexible IMO.)

(slightly edited my post for clarity)

Thanks armando for this crystal clear explanation. Also related is this book page: 1. Grid Display Modes

Many thanks Armando! I am starting to get the idea but a while to go... :)

So in the getting started guide it says 'items are items' but in a sense there's a difference between 'root' items and subitems. But, this is only true for column filters- creating a new grid with a source field treats all items equally, but shows context parents where desired. I guess the question from that is:

How is the notion of parents stored in the data fields? I see it in the properties pane, how does the data there relate to field values? Given you can have more than one parent, how does IQ decide which parent is shown as the context parent in other grids?

Probably a faulty question anyway until I understand the model more, so if its nonsense please disregard! :)

>How is the notion of parents stored in the data fields?

It isn't. an IQBase has 6 main tables:

Items
Field definitions
Field values
Grids
Forms
Item links

So items exist in the Items table, independent of what's in the Item links table. In fact, the current parent-child links are just one type of links. More will be implemented. The next type is the relational link. Father - son, Employer - employee. These will be more flexible and more powerful since the link will have a qualifier. Right now, the parent-child do not have qualifiers, it is an item - sub-item link, nothing more. These new links will be displayed in a separate pane.

>Given you can have more than one parent, how does IQ decide which parent is shown as the context parent

By default, the first parent is the main parent. Right-click on a parent in the properties pane and select "Set as main parent" to change it.

Ah! I see. That link qualifier idea bends my brain. This program reminds me of the old TV show 'Twin Peaks'. Just when it all starts to sit together another fact emerges which makes you see all the previous stuff from a new angle again!

That's a good thing though - the flexibility and power of IQ is amazing. I'm going to start using it full time over the next week or so (assuming we aren't expecting the IQBase format to change anytime soon without a converter of some sort!). That should get the structure clearer in my skull... :)

>I'm going to start using it full time over the next week or so

Great!

>(assuming we aren't expecting the IQBase format to change anytime soon without a converter of some sort!)

The format is quite stable and when changes are made, it is transparent to users.

Thanks! I'm looking forward to it... :)

[quote=Armando] . . I use a different strategy than most users here as far as how I use one specific source field for each grid, in most of my grids, that is propagated to all items, in addition to another source field that "characterizes" the items that are actually of a certain "grid type"/category -- I think I described my method in another thread . .[/quote]

any ideas where that thread/post is Armando ? (or clues of what to search for?)

thanks,

Tom

Just saw your post. Sorry...

Look at : Flags for grid visibility - how to approach this?

and Copying items -- but that one is almost the same, just older...