jmespath/chat

Ryan Merolle

@ryanmerolle

but I want in this case, 2 lists (or however many dictionaries that I find in the above list)

thanks as always

nevermind!

[].endpoints.[*].ip does the trick...answered my own question

Doqnach

@Doqnach

@ryanmerolle [].endpoints[*].ip gets you in the right direction

ow you fixed it already :-P

Ryan Merolle

@ryanmerolle

thanks @Doqnach

Doqnach

@Doqnach

think you have a . too much

Ryan Merolle

@ryanmerolle

yea that was a typo on chat

Mitch Schooler

@MitchSchooler_twitter

Hi, I'm trying to use jmespath within a Tavern API test executed by pytest and I don't seem to be able to use the search() function. Other jmespath functions are working fine but for some reason I keep getting "jmespath.exceptions.UnknownFunctionError: Unknown function: search()" when trying to use the search function. Any idea why this one would not be found when other jmespath functionality is working as expected?

I'm trying to do something like:

          function: tavern.testutils.helpers:validate_content
          extra_kwargs:
            comparisons:
              - jmespath: "search(members, @)"

Chris Armstrong

@chris-armstrong

jmespath itself doesn't have a search() function - this is typically the API name for calling JMESPath with an object and expression.

You probably just want a filter expression i.e. array[?some_property == 'value1']

Mitch Schooler

@MitchSchooler_twitter

Okay, thanks for that clarification. I'm sure I'm just missing something super obvious here, but I thought search() was how you accessed the value of a particular key in a map. What is the proper way to do this? For example, how do I get "bar" out of this map the way that this search() example shows without search()?

search(foo, {"foo": "bar"}) -> "bar"

Chris Armstrong

@chris-armstrong

bar - literally just the property name

Mitch Schooler

@MitchSchooler_twitter

In my case I only know the expected key (foo) and its value could vary. How do I get the value out of the map using the foo key?

Chris Armstrong

@chris-armstrong

sorry I meant foo

Mitch Schooler

@MitchSchooler_twitter

Oh, so like: @[foo] ?

Chris Armstrong

@chris-armstrong

no just foo or @.foo

Mitch Schooler

@MitchSchooler_twitter

Gotcha, I'll give that a try. Thanks!

Chris Armstrong

@chris-armstrong

@[foo] gives you ['bar']

Mitch Schooler

@MitchSchooler_twitter

I see. So how would I get "ids" from this nested example? Would it just be @.members.ids and it should give me back ["id1", "id2"] ?

{
    "members": {
        "ids": [
            "id1",
            "id2"
        ]
    }
}

Chris Armstrong

@chris-armstrong

yep

sinusx

@sinusx_gitlab

Hi! Week ago I have found online tool where I can make query by jmespath language, now I can not find it again. Please provide link if you know?

jdevillard

@jdevillard

Hello you’ve different tool available

The site jmespath.org with de javascript impl

And https://jmespathtest.azurewebsites.net with dotnet impl

sinusx

@sinusx_gitlab

@jdevillard Thank you!!!

broskees

@broskees

Hey guys!

Discovered JMESpath a few days ago

It's helped me immensely already

Quick question to whoever is available

With a JSON object like the following

[
    {
        "type": {
            text: "company"
        },
        "data": {
            "name": "Company A",
            "logo": "company_a.png",
            "divisions": "Division A, Division B, Division C",
            "homepage": "company-a.com"
        }
    },
    {
        "type": {
            text: "company"
        },
        "data": {
            "name": "Company B",
            "logo": "company_b.png",
            "divisions": "Division A, Division B, Division C",
            "homepage": "company-b.com"
        }
    },
    {
        "type": {
            text: "employee"
        },
        "data": {
            "name": "John Smith",
            "dob": "1/16/1990",
            "position": "CEO"
        }
    },
    {
        "type": {
            text: "company"
        },
        "data": {
            "name": "Company C",
            "logo": "company_c.png",
            "divisions": "Division A, Division B, Division C",
            "homepage": "company-c.com"
        }
    },
    {
        "type": {
            text: "employee"
        },
        "data": {
            "name": "Oprah",
            "dob": "1/16/1990",
            "position": "TV Personality"
        }
    }
]

Is there a JMESpath query I can use to reformat the above string to display more like this?

{
    companies: [
        {
            "type": {
                text: "company"
            },
            "data": {
                "name": "Company A",
                "logo": "company_a.png",
                "divisions": "Division A, Division B, Division C",
                "homepage": "company-a.com"
            }
        },
        {
            "type": {
                text: "company"
            },
            "data": {
                "name": "Company B",
                "logo": "company_b.png",
                "divisions": "Division A, Division B, Division C",
                "homepage": "company-b.com"
            }
        },
        {
            "type": {
                text: "company"
            },
            "data": {
                "name": "Company C",
                "logo": "company_c.png",
                "divisions": "Division A, Division B, Division C",
                "homepage": "company-c.com"
            }
        }
    ],
    employees: [
        {
            "type": {
                text: "employee"
            },
            "data": {
                "name": "John Smith",
                "dob": "1/16/1990",
                "position": "CEO"
            }
        },
        {
            "type": {
                text: "employee"
            },
            "data": {
                "name": "Oprah",
                "dob": "1/16/1990",
                "position": "TV Personality"
            }
        }
    ]
}

broskees

@broskees

Basically I want to filter the objects by the type.text, and display the top level object containing that type.

I've been playing with the contains function, which does allow me to filter, however the return result is only the current node, I need to return the containing parent node.

Is this possible with JMESpath?

Chris Armstrong

@chris-armstrong

It seems it is: { companies: [?type.text=='company'], employees: [?type.text=='employee'] }

Evgeny Zislis

@kesor

Any way to query a field that is sometimes a single and sometimes a list? Like sometimes it is { "item": { "a" : "b" } } but sometimes it is { "item": [ {"a":"b"}, {"c":"d"} ] } and I would like to always convert it to an array/list when it is in the first representation

Ah, looks like I found a way ... first always convert to array, then flatten. Like this [SomePathWhereItemsAre.item][] takes care of a single item, and of multiple items - always returns a list (of one or more).

broskees

@broskees

@chris-armstrong That worked!! How would you do it if the beginning string was an object, like so?

{
    "0": {
        "type": {
            "text": "company"
        },
        "data": {
            "name": "Company A",
            "logo": "company_a.png",
            "divisions": "Division A, Division B, Division C",
            "homepage": "company-a.com"
        }
    },
    "1": {
        "type": {
            "text": "company"
        },
        "data": {
            "name": "Company B",
            "logo": "company_b.png",
            "divisions": "Division A, Division B, Division C",
            "homepage": "company-b.com"
        }
    },
    "2": {
        "type": {
            "text": "employee"
        },
        "data": {
            "name": "John Smith",
            "dob": "1/16/1990",
            "position": "CEO"
        }
    },
    "3": {
        "type": {
            "text": "company"
        },
        "data": {
            "name": "Company C",
            "logo": "company_c.png",
            "divisions": "Division A, Division B, Division C",
            "homepage": "company-c.com"
        }
    },
    "4": {
        "type": {
            "text": "employee"
        },
        "data": {
            "name": "Oprah",
            "dob": "1/16/1990",
            "position": "TV Personality"
        }
    }
}

Chris Armstrong

@chris-armstrong

I think you need an extra [] operator e.g. { companies: [][?type.text=='company'], employees: [][?type.text=='employee'] }

broskees

@broskees

Hm, that doesn't seem to be working for me. Last question

Just returns { companies: null, employees: null }

Chris Armstrong

@chris-armstrong

oh I thought the flatten operator would just give the values.
try values(@) | { companies: [?type.text=='company'], employees: [?type.text=='employee'] }

kguggulotu

@kguggulotu

I'd like to use jmespath to transform json from:

{
    "order": {
        "order_type": "online",
        "item": [
            {
                "item_nbr": 1,
                "item_type": "so",
                "mlog": [
                    {
                        "id":"1",
                        "timestamp":"2019-12-13"
                    },
                    {
                        "id":"2",
                        "timestamp":"2019-12-13"
                    }
                ]
            },
            {
                "item_nbr": 2,
                "item_type": "so",
                "mlog": [
                    {
                        "id":"1",
                        "timestamp":"2019-12-13"
                    },
                    {
                        "id":"3",
                        "timestamp":"2019-12-13"
                    }
                ]
            }
        ],
        "mlog": [
            {
                "id":"5",
                "timestamp":"2019-12-13"
            },
            {
                "id":"6",
                "timestamp":"2019-12-13"
            }
        ]
    }
}

[
    {
        "order": {
            "order_type": "online",
            "item": [
                {
                    "item_nbr": 1,
                    "item_type": "so",
                    "mlog": [
                        {
                            "id": "1",
                            "timestamp": "2019-12-13"
                        },
                        {
                            "id": "2",
                            "timestamp": "2019-12-13"
                        }
                    ]
                }
            ],
            "mlog": [
                {
                    "id": "5",
                    "timestamp": "2019-12-13"
                },
                {
                    "id": "6",
                    "timestamp": "2019-12-13"
                }
            ]
        }
    },
    {
        "order": {
            "order_type": "online",
            "item": [
                {
                    "item_nbr": 2,
                    "item_type": "so",
                    "mlog": [
                        {
                            "id": "1",
                            "timestamp": "2019-12-13"
                        },
                        {
                            "id": "3",
                            "timestamp": "2019-12-13"
                        }
                    ]
                }
            ],
            "mlog": [
                {
                    "id": "5",
                    "timestamp": "2019-12-13"
                },
                {
                    "id": "6",
                    "timestamp": "2019-12-13"
                }
            ]
        }
    },
    {
        "order": {
            "order_type": "online",
            "mlog": [
                {
                    "id": "5",
                    "timestamp": "2019-12-13"
                },
                {
                    "id": "6",
                    "timestamp": "2019-12-13"
                }
            ]
        }
    }
]

Can someone help me with the query ?

Chris Armstrong

@chris-armstrong

It looks like a fairly complicated transformation - you might want to break it down

Rajat Taneja

@rajataneja101

Hi all, lets say i have a json like following: [
{
"Name": "JsonOneName",
"Description": "JSONONEDESCRIPTION"
},{
"Name": "JsonTwoName",
"Description": "JsonOneDescription"
}
]
Now the problem which i am facing is, in the above json both the blocks have a same json block with same description just the casing is different so how can i run my json query without the case sensitivity even though the description is same??

alexanderinsa

@alexanderinsa

I'm new to working with jmespath and I was wondering if there was a way to select multiple dictionary keys based on if that key contains a string? So like {a:blah, aa:blah, c:notblah} I would like to flatten the results of the keys that contain an a?

Where communities thrive

People

Repo info

Activity