I would like to pass a nested key to jq in an arg. Using a single --arg key $mypath fails.
Looking at other questions it seems split is the correct method, but I cant find a basic example that works.
mypath="level1.level2.level3"
//working hardcode
jq '.level1.level2.level3' $myfile
//variations tried
jq --arg nest "${mypath}" '.[$nest]' $myfile
jq --arg nest "${mypath}" '.[$nest]|split(".")' $myfile
jq --arg nest "${mypath}" '.|split(".")|[$nest]' $myfile
You need getpath/1 for this. The method takes an array of paths to a particular node and returns the value present at it
jq --arg nest "$mypath" 'getpath($nest|split("."))'
Related
I have a list of country codes like FR, IT, DE and have been trying to figure out how to use this in a select statement. I was doing something like
cat stuff | jq -c '.[]| select(.country_iso3166_alpha2 == "US")'
But then my list grew to a large number of countries I want to match on. So I tried using IN since I'm using jq 1.6 and did something like this:
eu=("FR", "IT"); cat stuff | jq -c '.[]| select(.country_iso3166_alpha2 IN($eu)'
I've been reading the docs and looking at the cookbook but it's not making any sense to me. Thanks!
You can use --argjson to pass the list to jq and IN to select the matching entries.
jq -c --argjson eu '["FR", "IT"]' '.[]| select(.country_iso3166_alpha2 | IN($eu[]))' <stuff
Broken out to show the individual parts:
jq -c \
--argjson eu '["FR", "IT"]' \
'.[]| select(.country_iso3166_alpha2 | IN($eu[]))' \
<stuff
invoke jq with compact output
pass in the list of countries as a json array named "eu"
select using the IN operator, unpacking $eu[] to get its values
redirect the input file into jq
Unfortunately, jq does not understand bash arrays.
Things would probably be simplest if you can arrange to have your shell variable be a string representing a JSON array:
eu='["FR", "IT"]'
jq -n --argjson eu "$eu" '$eu'
Or if you prefer:
eu='["FR", "IT"]' jq -n 'env.eu | fromjson'
Another possibility would be to write your bash array into a file, and then slurp the file.
There are also many variants of the above....
I have a number of JSON files that look like this:
{
"property": "value1",
...
}
What I want is an output file that looks like this:
{
"<filename1>": "<value1>",
"<filename2>": "<value2>",
"<filename3>": "<value3>",
...
}
This can be achieved with two jq invocations and a shell pipe:
jq '{(input_filename):.property}' * | jq -s add
However, I was wondering whether this is possible with a single jq invocation (or any other simpler way).
I'm currently using jq version 1.5-1 in case it's relevant.
Use inputs in combination with the -n option to sequentially access all input files.
In direct analogy, you could just create the array that would have been created by the -s option using [inputs], and then add up the items as you did before:
jq -n '[inputs | {(input_filename): .property}] | add' *
But in a more straightforward way, you could employ reduce to iteratively build up your result object:
jq -n 'reduce inputs as $in ({}; .[input_filename] = $in.property)' *
I'm trying to out data with jq request
"{\"#timestamp\":\"2019-03-13T00:11:03.123Y\"
I'm typing: cat myfile.txt | jq fromjson.timestamp
But it's incorrect
The pipeline you're looking for is
jq 'fromjson | .["#timestamp"]'
You cannot use the .foo syntax to access fields with special characters such as "#".
I am trying to run a jq query on a windows machine and it extracts values from output on a separate line
jq -r .Accounts[].Id
Output
204359864429
224271824096
282276286062
210394168456
090161402717
How do I run the jq query so that it combines the output on a single line separated by space
This is what I need-
204359864429 224271824096 282276286062 210394168456 090161402717
Any help would be appreciated.
The usual way would be to use the #csv or #tsv operators to convert the result in the CSV or tab-delimited format. These operators need the result to be contained in an array. For your case also to have a single space delimit, we can do a simple join(" ") operation
jq -r '[.Accounts[].Id]|join(" ")'
You can use the #sh formatter:
jq -r ".Accounts[].Id | #sh"
From the jq docs:
The input is escaped suitable for use in a command-line for a POSIX shell. If the input is an array, the output will be a series of space-separated strings.
Reference:
https://stedolan.github.io/jq/manual/#Basicfilters
At first I thought the join() solution above did not work. Then I realized that I was "overfeeding" the join() filter, causing it to fail because I was providing more than a simple array as input. I had concatenated several filters with , and failed to limit the scope of my join().
Did not work:
jq -r \
'.ansible_facts |
.ansible_hostname,
.ansible_all_ipv4_addresses | join(" "),
.ansible_local."aws-public-ipv4".publicIP'
This gave the error,
jq: error (at <stdin>:0): Cannot iterate over string ("hostone")
because jq was attempting to "consume" not only ansible_all_ipv4_addresses but also the output of the preceding ansible_hostname filter (I am not certain why this is or whether it was even intended by the author of jq).
Does work:
jq -r \
'.ansible_facts |
.ansible_hostname,
(.ansible_all_ipv4_addresses | join(" ")),
.ansible_local."aws-public-ipv4".publicIP'
Here, I restrict join() to .ansible_all_ipv4_addresses only (ansible_all_ipv4_addresses is an array of IP addresses I wish to translate into a single, space-separated string).
P.S.: I found that the #sh filter produces space-separated output as desired, but in addition delimits each output item in single quotes.
P.P.S.:
Here was my workaround, until I discovered that join() works just the same as it when used properly (see above):
jq -r '.Accounts[].Id | #tsv | sub("\t";" ";"g")'
Explanation: the #tsv filter produces Tab Separated Values, then the sub() filter substitutes tabs with spaces, globally.
If you'd like jq to escape dashes, you need to put your key between square brackets like this;
jq '.["key-key"]'
and apart from that, if you'd like to include a variable in jq, you need to use double quotes instead of single quotes;
jq "."${var[i+1]}""
but my variable contains dash in it and in this case, I've tried to merge the 2 examples above but it didn't work;
var=key-key
jq ".["${var[i+1]}"]."key""
how can I get this work?
Update:
This is the final script, which I've forgot to mention;
declare -a var=(
"key-key"
"key2-key2"
"key3-key3"
)
for ((i=0; i<${#var[#]})); do
curl -s "url" | jq ".["${var}"]."something""
done
To have double-quotes in a jq command you've enclosed in double-quotes, you'd escape them with a backslash :
jq ".[\"key-key\"]"
Another problem with your final command is that ${var[i+1]} expands to the empty string, because this syntax is used to index elements of an array, and you previously defined var as a simple string.
A better way to work with variables in jq commands is to define them through the --arg name value option, after which you can refer to them with $foo in a single-quotes enclosed command :
jq --arg keyName key-key '.[$keyName]'
To fix the code included in the update, I would use the following :
declare -a var=(
"key-key"
"key2-key2"
"key3-key3"
)
json=$(curl -s "url")
for searchedKey in "${var[#]}"; do
echo $json | jq --arg keyName $searchedKey '.[$keyName].something'
done