This thread will demonstrate the new nb204 anonymous procedure blocks in a sample expense tracker app. It will also show
Sample run:
The source, for those who just want the app:
monthly_category_totals.aia (13.7 KB)
Details and code to come in subsequent posts in this thread ...
Data storage and entry:
All data is kept in TinyDB, one tag per monthly expense entry.
The top of the app display is for data entry:
The Designer components:
Screen Initialization:
The Enter button is disabled until a category has been chosen.
Default current year and month are loaded, for convenience.
A procedure is called to refresh the data summaries (discussed later.)
The TinyDB tag structure:
I separate the parts of the TinyDB tag with '|' to make it easy to extract them later:
The Enter button builds a tag for TinyDB, and adds the amount to the existing amount:
Choosing Categories:
For flexibility, category options are chosen from the data base, with the option of selecting "[+]" to add a new expense category.
We need a distinct function to extract only distinct values from a list:
We also need a procedure to return only the TinyDB tags that are entries in our little data base:
(In this app, we use only one Namespace, and no other compound TinyDB tags, so this should suffice to return only the TinyDB tags for expenses. If you are testing in the Companion, you may need to change the TinyDB NameSpace to avoid other apps' data.)
Adding new Categories during data entry:
I use the "[+]" selection for adding new data, through a special purpose Notifier:
When we receive a new category value, we add it to the list of distinct existing categories from our TinyDB tags, and set it as our selected category.
Selecting a new Month is simple, since they have been pre-loaded in the Designer:
Next post: Filtering data.
Filters are kept as a dictionary, with column names ("Category", "Year", "Month", "Year.Month") as keys. The value under each key is the list of acceptable filter values for that column. If the disctionary is empty, everything is acceptable. If there are multiple column filters, each tag must satisfy all filters.
Designer:
Filters are presented in a ListView lvwFilters.
They can be added and removed (one filter column and value at a time) using two List Pickers and a Button.
Sample filter choices:
The nb204 Anonymous Procedures, at work:
This is where the column names meet the code to extract their values from TinyDB tags:
This dictionary is where the descriptive text names of the columns meet the code to extract that column value from a tag.
This code would have had to been coded as an if/then/elseif ladder before nb204.
Now the keys of this dictionary are available to be loaded into List Picker Elements to help the user configure his filters.
This is the filter dictionary we will be loading up with filters:
Using the new nb204 anonymous procedure blocks, we can look up and invoke the procedure to extract any requested column from a tag:
After a filter column is chosen, we use that column choice to extract all distinct values of that column from our TinyDB tags, and present them in the adjacent List Picker:
When a filter column value is chosen, we stage it for the add/remove button:
The Button to add and remove filter values needs both List Picker Selections, and reverses what it finds for this combination of column and value in the filters dictionary.
When the last filter value for a column is removed, that column name is removed from the filter dictionary.
The filter dictionary by itself is not enough.
We need a procedure to walk the list of TinyDB tags and return only the ones that pass all filters in the dictionary:
The reduce list block and the Make New Filtered List block are from nb191's advanced list blocks collection.
Here's the procedure to refresh the summaries so far:
Summary List Picker handling in next post:
For debugging and data correction, here are two List Pickers:
Next post: grouping
The grouping setup controls are after the filter controls in the display:
The grouping is done from a list of column names.
The List Picker chooses a column name to add or remove from the list of column names.
The ListView shows the grouping order, initially empty.
An empty grouping order provides a detailed list in the Group Summaries ListView.
Sample run: filtering by year, and grouping by category:
Sample run: grouping by category and year combination:
The group summarization is done in a dictionary, with each key derived from applying the group column names to each tag, and the key value is kept as the running sum of the matching tags' TinyDB values.
Assigning a group key to a tag:
The procedure to fill the group summaries:
That completes the blocks walkthru for this sample app.
Next post: Afterthoughts on Anonymous procedures
Afterthoughts:
@Taifun asks
This procedure
would have to be replaced with
I would still need a list of column names as in the keys of
to load some List Picker elements lists, but now I would need to make sure there are no typoes between the two collections of column names.
This is my first use of anonymous procedures.
I imagine them useful for apps like calculators and interpreters, where you have to map operation names '+', '-',... into code to apply those operations.
I open this thread for further discussion.
