Selenium: How to identify the button WebElement
By.xpath("//button[contains(.,'Add Strategy')]") By.xpath("//button[contains(.,'Submit')]")
Notice that it is same as:
By.xpath("//button[contains(text(),'Add Strategy')]") By.xpath("//button[contains(text(),'Submit')]")
Based on the html provided this is the best answer we could give. However, if more html were provided, we could probably find a better way to do it. Remember that xpath is one of the slower methods of finding items in selenium. Also at the risk of sounding like a broken record, I prefer not to use the inner text to identify elements since that gets localized and your automation would not run in localized builds, and the text is much more likely to change than an actual unique identifier in the markup.
The xpath examples here will work - but you could also use a css selector to match inside the contents of the "onclick" attribute.
In this case I've used the "starts with" attribute selector, but you could use the "contains" selector instead (
yes, as Tarun says, you can use xpath, but in our project, it's usually a required action for developer to add IDs to each elements. because,sometimes the UI or the text may change, then you had to modify your test script.
Tests of this sort should test the app the way a user would use it -- that is, they should test as much as possible based on UI, not element IDs. Therefore, they should find buttons based on text, not ID. Yes, that will require changing the test when the display text changes, but that's a *good* thing -- since the tests ensure that the UI is as desired, a UI change *should* break them.
I respectfully disagree with Marnen Laibow-Koser's statement. Testing based on the literal UI text leads to maintenance nightmares (been there, done that, it sucks). Validation of the text of a button, link, etc. should be a distinct test, not as a side effect of testing functionality. Finding objects on the page by literal text is also makes testing of pages translated to another language much more difficult.
This isn't about validation of the text, quite; rather, it's making sure the control is where the user can find and understand it. Remember, text is a UI element; classes are not. I know it may seem easier not to use the literal text, but that just leads to UI tests that don't actually test the UI.