I need to process a large json file, in which I need to change hyphens to underscore in values of a specific key, for example within the file I would change:
{
"key": ["test-value-1", "value-2"],
"anotherKey": "unrelated-value"
}
to:
{
"key": ["test_value_1", "value_2"],
"anotherKey": "unrelated-value"
}
without modifying the values of other keys. I think the problem can be generalised to replacing all the occurrences of a specific character between delimiters, in my case:
- the character to be replaced is
-, - the replacement is
_, - the delimiters are
"key": [and].
How can I achieve that using sed?
Note: I am using a macos, not sure if that matters.
So far I have managed to replace the entire value list with:
sed -E 's/"key": [(.*)]/"key": ["replaced_value"]/'
for example:
$ echo '{"key": ["test-value-1", "value-2"], "anotherKey": "unrelated-value"}' | sed -E 's/"key": [(.*)]/"key": ["replaced_value"]/'
{"key": ["replaced_value"], "anotherKey": "unrelated-value"}
I am not sure how to further operate on the value captured with the regular expression to replace only specific character within it.
3
You can just let sed replace all - with _ where the line contains "key"::
sed '/"key":/ s/-/_/g' /tmp/json
But using jq this is much more safer / easier with gsub:
jq '.key[] |= gsub("-"; "_")' /tmp/json
Online Demo
Both produce:
{
"key": ["test_value_1", "value_2"],
"anotherKey": "unrelated-value"
}
5