Using the textvariable attribute when creating a combobox in tkinter seems completely useless. Can someone please explain what the purpose is? I looked in the Tcl documentation and it says textvariable is used to set a default value, but it looks like in tkinter you would just use the .set method to do that.
Example showing what I mean:
This doesn't work...
from Tkinter import *
import ttk
master = Tk()
test = StringVar()
country = ttk.Combobox(master, textvariable=test)
country['values'] = ('USA', 'Canada', 'Australia')
country.pack()
# This does not set a default value...
test="hello"
mainloop()
This does work.
from Tkinter import *
import ttk
master = Tk()
country = ttk.Combobox(master)
country['values'] = ('USA', 'Canada', 'Australia')
country.pack()
# This does set a default value.
country.set("hello")
mainloop()
If you are supposed to just use the .set and .get methods, what is the point of assigning anything to textvariable? Every example online seems to use textvariable, but why? It seems completely pointless.
Since Python has no type safety, you're overwriting the reference to the StringVar object with a string. To set the value, call the set method:
test = StringVar()
country = ttk.Combobox(master, textvariable=test)
#...
test.set("hello")
Under normal circumstances there is no reason to use a StringVar. I have no idea why most tutorials show it. It adds overhead but provides no extra value. As you observe, you can directly get and set the value of the combobox via the combobox object itself.
The advantage to using a StringVar comes when you want to either a) have two widgets share the same variable so that one is updated when the other is changed, or b) attach one or more traces to the StringVar. Both of those are uncommon, but are sometimes quite useful.
As Squall pointed out, to change the StringVar value you need to use its set() method.
On the other hand, the purpose of Tk Variables is to keep track of the values of the widgets they are connected to. The reason why you should prefer using a StringVar over reading the widget's value directly is that the latter is thread-safe.
Thread safety is a bit of a thorny topic for tkinter, but technically you should not be allowed to make modification to GUI elements in a thread different from the one that created the window (i.e. the original thread blocked in mainloop() plus the callbacks that you defined).
If you have an extra background thread that elaborates data and changes the UI accordingly using the country.set("hello") instruction might crash your application. Some systems forbid this entirely.
Meanwhile, setting the value in a StringVar is perfectly safe.
If you never use other threads and can work from callbacks alone, you will never face this problem and you might consider StringVar as a useless overhead. However, you will need to use variables for any kind of serious work.
This was my first go at this method and I found myself here as well.
LOL I was curious as to why I must do that. I agreed that the .set method works a lot nicer in the code block. It's esthetically pleasing. It's just that sensible default.
I tried the above codes both failed. Language deprecation?
import tkinter as tk
from tkinter import ttk
master = tk.Tk()
country = ttk.Combobox(master)
country['values'] = ('USA', 'Canada', 'Australia')
country.pack()
This does set a default value.
country.set("hello")
master.mainloop()
Related
This question already has an answer here:
Python tkinter listbox bind on <Button-1> only works on second click
(1 answer)
Closed 1 year ago.
I was creating simple listbox containing numbers from 0 to 9. I wanted to print number when it get clicked so i bind list box with Button-1. Im facing problem that is when ever i select any number and try to get its location using list_box.curselection() it does not print any thing(return empty tuple), if i click on on any other number then it print previous selected number. I want to get current selected number.
from tkinter import *
root = Tk()
root.title("test listbox")
list_box = Listbox(root)
list_box.pack()
for i in range(0,10):
list_box.insert("end",i)
def def_fun(event):
print(list_box.curselection())
list_box.bind("<Button-1>",def_fun)
root.mainloop()
You don't have to bind to <Button-1> or anything, there is a virtual event with Listbox that you can use here:
def def_fun(event):
print(event.widget.curselection()) # The widget that triggers the event is event.widget
list_box.bind("<<ListboxSelect>>",def_fun) # Gets triggered each time something is selected
Just in case you are wondering why the Button-1 did not work, it is because there is a delay, the delay might be due to binding order, you can read more about it here but here is a gist:
In the default case, your binding on <Key> happens before the class binding, and it is the class binding where the text is actually inserted into the widget. That is why your binding always seems to be one character behind.
Change the binding to releasing mouse button, this will also be more user friendly (for example if they accidentally clicked on a selection they didn't want to select, they can move their mouse to a one they want and only releasing will call the function):
from tkinter import Tk, Listbox
def def_fun(event=None):
print(list_box.curselection())
root = Tk()
root.title("test listbox")
list_box = Listbox(root)
list_box.pack()
for i in range(0, 10):
list_box.insert("end", i)
list_box.bind("<ButtonRelease-1>", def_fun)
root.mainloop()
Another option if you want to call the function on select is either use #CoolCloud answer or you can also set a delay like this (although it will most certainly work in 99.9% of cases, there might be a case where it doesn't):
list_box.bind("<Button-1>", lambda e: root.after(10, def_fun))
The reason is that .curselection() gets the current selection but Button-1 is triggered before anything gets selected so it will print the previous selection because that is what was selected before and where the current selection is now, and then immediately after this, it will move the current selection to the item you clicked.
Important (because it may cause hard-to-debug issues):
I strongly advise against using wildcard (*) when importing something, You should either import what You need, e.g. from module import Class1, func_1, var_2 and so on or import the whole module: import module then You can also use an alias: import module as md or sth like that, the point is that don't import everything unless You actually know what You are doing; name clashes are the issue.
Also:
I strongly suggest following PEP 8 - Style Guide for Python Code. Function and variable names should be in snake_case, class names in CapitalCase. Don't have space around = if it is used as a part of keyword argument (func(arg='value')) but use if it is used for assigning a value (variable = 'some value'). Have two blank lines around function and class declarations.
I have a main program which does some cool stuff and I am currently setting up a 'settings editor' to let the user change some GUI related stuff and default values. It reads values from a text file, which is read in correctly and saves them to a dictionary self.propertiesDict. Some of the options are on/off switches, so I use checkbuttons for them. What is puzzling me is following behavior: the code works perfectly fine, when I execute the settingsEditor.py (the script creating the settings window) directly. All the checkbuttons are set to active / True. However, when I include my settingsEditor in my main program and call it, it creates fine but all the checkbuttons show the wrong value: False. I read a lot of topics here to find an answer, but I think I avoided the most common errors:
I use the tk variables
tk variables are created and set prior to the buttons
variables are not only in local scope (prefixed self.)
As you can see, I tried with an IntVar and a BooleanVar, but neither is working correctly. Something else is strange, when I use ttk.checkbuttons, I get the issue described here. I use Visual Studio for debugging and I can't see any difference in the process when going trough line by line, except for the wrong display result. I am happy for any suggestion. Sorry for not providing a full MWE, I will do, if nobody can help me from this here.
settingsEditor.py
import tkinter as tk
from tkinter import ttk
...
class mySettingsEditor:
def __init__(self):
...
def createGUI(self):
# Show main options on startup on/off
self.showOptionsVar = tk.IntVar()
self.showOptionsVar.set(str2int(self.propertiesDict['showMainOptionsExpanded']))
print(self.showOptionsVar.get())
self.checkBtn1 = tk.Checkbutton(Frame, text='Main Options Section', variable=self.showOptionsVar)
self.checkBtn1.grid(column=0,row=2)
# Show main STL section on startup on/off
self.showMainSTLVar = tk.BooleanVar()
self.showMainSTLVar.set(str2bool(self.propertiesDict['showMainSTLSectionExpanded']))
print(self.showMainSTLVar.get())
self.checkBtn2 = tk.Checkbutton(Frame, text='Main STL Section', variable=self.showMainSTLVar)
self.checkBtn2.grid(column=0,row=3)
main.py
from settingsEditor import mySettingsEditor
...
settEditor = mySettingsEditor()
This is how it looks in the GUI when executed separately (terminal with print output to the left):
Thats the result when I add it in main.py. The boxes are unchecked, but .get() tells me the values are correctly assigned to the tk variables.
As suggested by jasonharper, switching to Toplevel() for the child windows fixed the issue. Thanks alot!
Is there any way to add options in an OptionMenu object without using the command flag?
I have seen many posts, such as this one, showing how the add_command will update/remove/add options to an already existing OptionMenu object. For example, this snippet of code removes all options in serialPortOptionMenu and repopulates the option menu with different options:
serialPortOptionMenu["menu"].delete(0, "end")
for serialPort in serialPortsArray:
serialPortOptionMenu["menu"].add_command(label=serialPort[1], command=lambda v=serialPort: serialPortFunc(v))
However, something like this seems to overwrite the original command flag I wrote when creating the OptionMenu object:
serialPortOptionMenuValue = Tkinter.StringVar(optionMenuFrame)
serialPortOptionMenuValue.set(serialPorts[0])
serialPortOptionMenu = Tkinter.OptionMenu(optionMenuFrame, serialPortOptionMenuValue, *serialPorts, command=lambda *args: callbackFuncWhenOptionMenuSelectsAnotherOption(*args))
serialPortOptionMenu.grid(row=3, column=0, columnspan=2, sticky="we")
As I'm sure many people are wondering why I am setting a command within an OptionMenu (weird I know), it is because I want a callback function to be called when the user picks a new option.
"What about the trace option"-Everyone...Yes, I am aware of this as well, but because I am reading/writing new values to the Tkinter variable in code without having the OptionMenu solely change it, using trace within Tkinter would not be an effective substitute for just tracking when the user selects a new option in the OptionMenu.
Why am I changing the Tkinter value in code and not having the OptionMenu do it? Because I thought it would be nice to modify the Tkinter value string with something like a ~ at the end of the string to signify something is happening behind the scenes that isnt completed yet, and only when it is completed will the ~ go away, hence the reading and writing to the value without having the OptionMenu solely change it.
What I am primarily interested in is if anyone knows of other ways to add options to an OptionMenu object without using
myMenu["menu"].add_command(..., command=...)
As it seems to be removing my original callback function.
Note: Having two OptionMenus and performing grid_remove/grid on them to hide/reveal them also crossed my mind, but just seems to messy.
There is no other way. An option menu is nothing more than a menubutton and a menu with a custom binding. You can create your own OptionMenu and add any methods you want.
I don't know if it will be useful to you anymore, but maybe somebody else will be looking for this as I was. You actually just need to put the command as a callback parameter in the command you use to add the item (I found tk._setit):
def refresh_dropdown():
# Reset var and delete all old options
var.set('')
dropdown['menu'].delete(0, 'end')
# Insert list of new options (tk._setit hooks them up to var)
new_choices = ['a', 'b', 'c', '...'] # you can make a dictionary[key_from_previous_dropdown.get()] as I did in my case
for choice in new_choices:
dropdown['menu'].add_command(label=choice, command=_setit(var, choice, new_command))
dropdown = OptionMenu(app, var, command=new_command)
I am new in Python and in tkinter so the question may seems naive: is it ok to create and place widgets at the same time if I don't need to change them?
It works but is it a good practice? And if not why?
An example of what I mean:
import tkinter as tk
window=tk.Tk()
tk.Label(window,text='Lost Label').pack()
window.mainloop()
To expand upon #Skynet's answer....
Whenever you do Widget(*args, **kwargs).pack() the pack() method returns None as would other geometry managers, so if you tried to assign this to a variable the variable would be None.
In this case then probably not, since you probably actually want to be storing the reference to the widget.
If you don't need a reference then there's not really a problem with it. As the other answer notes you don't need a definitve reference to every single widget in your GUI unless you plan to use this reference in some way. Unless I plan on changing the label text / modifying it in someway then I typically use your method to save some space. No need to write more code than you have to!
For example you're creating a Button widget.
btn = Button(blabla, text="Button1")
This returns a button object and if you need later to configure it or get information about it you can do it by through the btn variable.
But if you use something like btn = Button(blabla, text="Button1").pack() it returns None and not a button object so you won't be able to change anything about the button or get information about it later.
Another example is with the Entry widget
entry = Entry(blabla)
Using that later you can do entry.get() to get the text inside the entry
but you won't be able to do it if you use entry = Entry(blabla).pack() since it doesn't return an entry object, it just packs the widget and you won't be able to access it for later use.
There is nothing wrong with that approach and I have already seen it quite a few times. You don't have to keep a reference to every widget in your GUI.
I have an Entry widget on a simple calculator. The user can choose to enter an equation via the keypad. I was wondering if there was a way to detect a character(from the keypad in my case) being typed into the Entry widget. So, focus is on the widget, user presses '4', it comes up on the widget... can I detect this act, for basic purposes of logging the input?
Every time you press a key inside a Tkinter window, a Tkinter.Event instance is created. All you need to do is access that instance. Here is a simple script that demonstrates just how:
from Tkinter import Tk, Entry
root = Tk()
def click(key):
# print the key that was pressed
print key.char
entry = Entry()
entry.grid()
# Bind entry to any keypress
entry.bind("<Key>", click)
root.mainloop()
key (being a Tkinter.Event instance) contains many different attributes that can be used to get almost any type of data you want on the key that was pressed. I chose to use the .char attribute here, which will have the script print what each keypress is.
Yes. There are a few different ways to do this, in fact.
You can create a StringVar, attach it to the Entry, and trace it for changes; you can bind all of the relevant events; or you can add a validation command that fires at any of several different points in the sequence. They all do slightly different things.
When a user types 4, there's a key event with just the 4 in it (which doesn't let you distinguish whether the user was adding 4 to the end, or in the middle, or replacing a whole selected word, or…), and then a modification event is fired with the old text,* and then the "key" or "all" validation function is called with the (proposed) new text, and the variable is updated with the (accepted) new text (unless the validation function returned false, in which case the invalidcommand is called instead).
I don't know which one of those you want, so let's show all of them, and you can play around with them and pick the one you want.
import Tkinter as tk
root = tk.Tk()
def validate(newtext):
print('validate: {}'.format(newtext))
return True
vcmd = root.register(validate)
def key(event):
print('key: {}'.format(event.char))
def var(*args):
print('var: {} (args {})'.format(svar.get(), args))
svar = tk.StringVar()
svar.trace('w', var)
entry = tk.Entry(root,
textvariable=svar,
validate="key", validatecommand=(vcmd, '%P'))
entry.bind('<Key>', key)
entry.pack()
root.mainloop()
The syntax for variable trace callbacks is a bit complicated, and not that well documented in Tkinter; if you want to know what the first two arguments mean, you need to read the Tcl/Tk docs, and understand how Tkinter maps your particular StringVar to the Tcl name 'PY_VAR0'… Really, it's a lot easier to just build a separate function for each variable and mode you want to trace, and ignore the args.
The syntax for validation functions is even more complicated, and a lot more flexible than I've shown. For example, you can get the inserted text (which can be more than one character, in case of a paste operation), its position, and all kinds of other things… but none of this is described anywhere in the Tkinter docs, so you will need to go the Tcl/Tk docs. The most common thing you want is the proposed new text as the argument, and for that, use (vcmd, '%P').
Anyway, you should definitely play with doing a variety of different things and see what each mechanism gives you. Move the cursor around or select part of the string before typing, paste with the keyboard and with the mouse, drag and drop the selection, hit a variety of special keys, etc.
* I'm going to ignore this step, because it's different in different versions of Tk, and not very useful anyway. In cases where you really need a modified event, it's probably better to use a Text widget and bind <<Modified>>.
If you just need to do simple things without using trace module you can try
def objchangetext(self, textwidget):
print(textwidget.get()) #print text out to terminal
text1 = tk.Entry(tk.Tk())
text1.bind("<KeyRelease>", lambda event, arg=(0): objchangetext(text1))