DOMXPath is encoded into the webpage you’re capturing, and is a read only property. REcorder capture allows multiple properties for identification of the control captured. I wouldn’t recommend using only DOMXPath as the selection critera.

Can you provide more information with example and details for your use case?

Couple of things.

1) Once you capture an object you will have a list of object properties, if you expand this you can select and unselect the search criteria used for finding an option. (this will allow you to select html name ect.)

2) If your object is in a frame HTMLHasFrame = true then you also should select HTML FramePath or FrameDomPath to tell AA what frame you are selecting

3) You need to select the minimum number of object properties that will get you your unique object. This changes website to website and case by case. However, for example if you know that in the HTML your object has a unique ID you don’t need a dompath or path to find this object just the id. To be honest I disagree with Jon above I think it’s completely fine to use just DomXpath to find an object in certain scenarios. You just have to test and confirm that it’s working properly and AA can find the object consistently.  

• Just a general tip always unselect path, it’s auto selected for you when capturing an object and it’s the most useless, error invoking property. To this day I don’t know why it’s still included in the search criteria, maybe 1/100 objects actually need this and if an object does need it it’s probably more effective to use image recognition to find it anyway. (sorry rant over)