Many Ansible Modules return a set of objects. When these objects do not have a unique identifying attribute, the data is often returned as a list of dictionaries. For example, the built-in setup module returns mount points this way. (I omitted some attributes for better readability.) "ansible_mounts": [ { "device": "/dev/sda1", "fstype": "vfat", "mount": "/boot/efi", "size_available": 202100736, "size_total": 209469440 }, { "device": "/dev/mapper/rootvg-rootlv", "fstype": "xfs", "mount": "/" }, { "device": "/dev/mapper/rootvg-homelv", "fstype": "xfs", "mount": "/home" } ] I understand, that the device or mount entries do not necessarily need to be unique, so a list is the only structure that can always be constructed. But in my case, they are. So I would prefer this data structure instead. { "/boot/efi": { "device": "/dev/sda1", "fstype": "vfat", "size_available": 202100736, "size_total": 209469440 }, "/dev/mapper/rootvg-rootlv": { "fstype": "xfs", "mount": "/" }, "/dev/mapper/rootvg-homelv": { "fstype": "xfs", "mount": "/home" } } Is there a combination of built-in filters that can easily transform one into the other? A filter plugin doing this can be written in a few lines of Python, though I was unable to find one on Ansible Galaxy. So before I upload one myself I would like to know if I am overlooking an obvious built-in solution. Continue reading...
