Filter by values using ./jq - json

Question

Given the input of sizes:
[
{
"stock": 1,
"sales": 0,
"sizes": [
{
"countries": ["at", "be", "ch", "cy", "de", "ee", "es", "fi", "gr", "ie", "lu", "lv", "nl", "pl", "pt", "se", "si", "sk"],
"size": "EU 45,5"
},
{
"countries": ["it"],
"size": "EU 45,5"
},
{
"countries": ["fr"],
"size": "EU 45,5"
},
{
"countries": ["gb"],
"size": "EU 45,5"
}
]
}
]
I will like to get the same structure without the ones that countries hasn't "de" (Germany) and remove the field complete. Expected something like this:
[
{
"stock": 1,
"sizes": [
{
"size": "EU 45,5"
}
]
}
]
I tried this:
map(.sizes[] |= select(.countries | join(",") | contains("de"))) | map({ stock, sizes })
But the filter is not working properly, throwing jq: error (at <stdin>:48): Cannot iterate over null (null).
Tried has, in, contains, inside and nothing seems to work.
Also, how can I filter which field appears? With map({ stock, sizes }) countries still there. Can I do something like map({ stock, sizes: { size } })?

Answer 1

Here's a one-liner that answers your main question -- if you can't see how it works, try breaking it up into separate pieces:
map( .sizes |= map( select(.countries | index("de") ) | del(.countries) ))
Regarding the selection of fields, you can use del/1 as above, or sometimes simply using an expression such as {key1, key2} will do the trick. Consider also this function and the following example:
def query(queryobject):
with_entries( select( .key as $key | queryobject | has( $key ) ));
Example:
$ jq -c -n '{"a": 1, "b": null, "c":3} | query( {a,b,d} )'
{"a":1,"b":null}