flattening a json file using jq

I would like to flatten a json file in order to meet the needs of a legacy program. I would prefer to use jq for this as I know it is available to me.

I have a file something like this:

[
{
"field1":10,
"field2":20,
"msg":
  {
  "a":1,
  "b":2,
  "c":3
  }
}
]

For reasons, I need to pull the a and b fields from msg and flatten their names:

[
{
"field1":10,
"field2":20,
"msg.a":1,
"msg.b":2,
"msg":   # the entire msg structure is now optional; if it falls off that's ok
  {
  "a":1,
  "b":2,
  "c":3
  },
}, ...
]

How is the miracle accomplished?

>Solution :

This works fine with the example input:

map(. + (.msg | {"msg.a": .a, "msg.b": .b}))

This is more flexible if the actual task involves more keys or keys with longer names:

map(. + (.msg | {a, b} | with_entries(.key |= "msg.\(.)")))

And if you do this for several fields like msg, a function could help shorten the code:

def f(x): . as $in
| reduce path(x) as $p ({};
  . + {($p | join(".")): $in | getpath($p)}
);
map(. + f(.msg | .a, .b))