Question Int.Parse throws invalid format error, but the string seems to be a valid integer

Hi, I have a <string,string> dict, and some of the values contain an integer. Later, I'm trying to parse an int value like

Code (CSharp):

1. Int.Parse(myDict[key])

.
It throws a format error even though I'm sure it contains just a single number as the string, after checking with debugging and Debug.Log.
I'm collecting the numbers from a TMP_textfield. Could this part of the issue somehow?

Thanks!

 

Can you show an example of what should be a valid string (taken straight from the logs)?

 

Code (CSharp):

1. Debug.Log(spaceShipStats["ShipAmount"]);

prints

Code (CSharp):

1. 1
2. UnityEngine.Debug:Log (object)
3. BattleController:ConstructShips () (at Assets/Scripts/BattleController.cs:86)
4. UnityEngine.EventSystems.EventSystem:Update () (at Library/PackageCache/com.unity.ugui@1.0.0/Runtime/EventSystem/EventSystem.cs:501)

 

Well, if some of your values aren't integers, you should probably be using TryParse anyways. Try outputting it from here what you get before and after using it.

 

Well, int.Parse does not have any room for interpretation. When it says there's something wrong with the string, there is something wrong with it. You said you "printed it out", which indeed is the first step, but does not rule out any issues with the string. Leading or trailing spaces, control characters, null character of byte order marks would cause such an error. The next simple step is to also print out the Length of the string. Since you said (and the log seems to confirm) that the string only contains a single character, the Length should be 1. Any larger value already indicates that there are other characters involved.

Try adding

using System.Linq;

at the top and print this instead:

Code (CSharp):

1. string hexStr = str.Aggregate("", (a, c) => a + ((int)c).ToString("x4")+" ");

It will convert each character in the string into the corresponding hex code. That way you will see every character that there is and you can also identify what those characters actually are. If you can't interpret the output yourself, post it here.

 

I believe that Int.Parse tolerates whitespace -- but it's definitely possible that there're some weird invisible characters in there. Text is always such a joy to deal with...

 

Hi all.

The problem indeed seemed to be extra invisible characters in the string. When looking at the textInfo, the single digit contained 9 or so characters...

I obtained the string from the TMP_Text object inside a TMP_InputField - but directly using the InputField.text magically fixed the problem.

Thanks for the suggestions!

 

Yes, you're right, it seems that spaces, tabs and new line characters are allowed as leading and trailing characters. It even allows the null character as a trailing character. Though as a leading character it would still fail.

Well I guess it may have added some formatting rich text tags which are not displayed in the console, as the console can also interpret those That's why just printing out strings does not necessarily tell the whole story. I recently had a similar issue at my job (not in Unity). I was printing stuff into the browser console but it also interpreted / stripped certain character I did not expect to get stripped ^^.

 

There's a lot to this. The text field you got that from can be elided (Such as when you type a very long string that scrolls off the left), plus I think it has these weird UTF-8 inviso spaces jammed in for things, plus possibly an inverted thin space character for the caret.

Always get your data from the source, which in this case is InputField.text. Same applies for the TMPRo versions of InputField and TextMesh

I remember discovering this first-hand in 2014 and it baffled our entire team for a day until we nailed it down.

 

Oh! I completely missed that.

That'll do it.