I am attempting to use the EvaluateJsonPath processor in Nifi, and am having trouble with the jayway jsonpath syntax.
My object looks like the following:
{"text":"my stuff", "tags":["abc", "xyz", "beq"]}
I want to route messages based on the tags - I want everything containing "xyz" to be routed one way, and everything not containing it to be routed another way.
Using http://jsonpath.herokuapp.com/ I've been testing and trying to figure out the syntax to filter based on a json object containing an array of strings matching. I can match based on overt index (so $.[?(#.tags[1] =~ /xyz/i)] works just fine), but I can't guarantee the order or number of objects in the tags field.
Is there a way to do this in the jayway json module? I saw filter the Json according to string in an array in JSONPATH which I've tried, but it doesn't appear to work in the simulator above.
I do not know how to do this in one EvaluateJsonPath processor step. But it can certainly be done in a two-step process:
Use EvaluateJsonPath to filter "xyz" tags out of the tags array, using a JsonPath expression like $.tags[?(# =~ /xyz/i)] and setting the processors return-type to json so an array may be returned. This will result in ["xyz"] for a match and [] for non-matching files
Use RouteOnAttribute to route based on the resulting array, with an expression like ${matchingTags:toLower():contains('xyz')}.
It might also be worth considering evaluating the JSON as text against a regular expression to match the tag.
Related
Hi this is a simple json
{"one": 335, "two": 331,"three": 65}
i need the json path to evaluate if a property has a value, example if exists "two":331
i tried this but it doesn't match.
$.two==331
what is the right expression?
You can check if there is any property in the given document with value 331 with this jsonpath expression: $[?(#==331)] But this will result an output like [331]. There is no chance to get the name of the matching property by using jsonpath only. You have to embed your expression into a script which supports jsonpath execution. Python does. Newer .NET Core based solutions (written with C# 9 and above) look also pretty simple.
I am using Handlebars templates. If I want to insert a variable in an hbs file, I add it as {{var_name}}.
The issue is, I have a large object with its keys taking formats similar to:
{
"stats.name":"John",
"stats.performance.day":123,
"stats.performance.month":4567,
"company":"My LLC"
}
If I try to add these to the Handlebars file as {{company}}, {{stats.name}}, {{stats.performance.day}}, {{stats.performance.month}} then only the {{company}} will be displayed. All other values come out blank, I assume because "." is a special character in Handlebars.
My question is, is there a way to override this, or do I actually need to iterate through this object and change every "." to a "_" or something before passing it to Handlebars to make it work?
I would recommend renaming your properties. Dot Notation is one of the two ways to access object properties in JavaScript. Typical code would look like: obj.prop. If your object's keys are not valid identifiers, then you must use Bracket Notation: obj['my-prop']. Since a dot is a property accessor, it is very unusual to see one in a property name. If I were to come across code written as obj['stats.performance.date'], I would assume it was a mistake and that obj.stats.performance.date was intended.
With that said, Handlebars does support referencing properties that are not valid identifiers. This is called "segment-literal-notation", and it is as simple as wrapping square brackets around your identifier:
{{[stats.performance.day]}}
The documentation also states that:
JavaScript-style strings, " and ', may also be used vs. [ pairs.
So the following alternatives are valid:
{{"stats.performance.day"}}
{{'stats.performance.day'}}
JMeter path contents some dynamically generated value.
Eg
[HTTP Request]
[path-home/user?p=er3562]
This "p" value is dynamically generated.
I want to get this parameter value in the first HTTP request path.
This parameter value should pass through each HTTP request path.
I am new to JMeter. Please help me to solve this?
You need to extract it somehow and convert into a JMeter Variable for later reuse. JMeter provides several PostProcessors for extracting data from different responses types, in your case the most suitable one will be Regular Expression Extractor.
Add Regular Expression Extractor as a child of the request which returns that path-home/user?p=er3562 value
Configure it as follows:
Apply to: depending on where the "interesting" value lives, the most "safe" setting is Main sample and sub-samples
Field to check: depending on where the "interesting" value comes from, in the majority of cases it's Body but in your case it may be i.e. URL
Reference Name: anything meaningful, it is JMeter Variable name, if JMeter finds anything it will store the result in a variable named accordingly to this field. I.e. path-home
Regular Expression: Perl5-style regular expression, in your case it would be something like: path-home/user\?p=er(\d+)
Template: if you're looking to extract a single value it will be $1$
Refer extracted value as ${path-home} where required.
References and tips:
You can use Debug Sampler and View Results Tree listener combination to view JMeter Variable names and test regular expressions against actual response
Using RegEx (Regular Expression Extractor) With JMeter
Perl 5 Regex Cheat sheet
JMeter Regular Expressions
By using correlation concept ,
regular expression extractor concept & try this key
p=er(.*?)
I have the following objective. I want to find, which documents contain my data when executing any kind of Xquery or XPath. In other words, I need every document that is providing me the result data for a given query. I try to do this in eXist-db environment, but I suppose there should be something on Xquery level.
I found op:context-document() operator which seems to have functionality I want, yet, as an operator it is not available for me. fn:document-uri also does not do the trick, as its $arg must be a document node, otherwise it returns an empty sequence.
Do you have any idea in mind? All the assistance is highly appreciated.
fn:base-uri() may help; it returns the base URI property of a node:
for $d in doc('....')/your[query]
return base-uri($d)
You can also use it to filter your documents for specific types:
collection('/path/to/documents')[ends-with(base-uri(), '.xml')]
Use the standard XPath / XQuery function collection() .
For example, using Saxon:
collection('file:///a/b/c/d?select=*.xml')[yourBooleanExpression]
selects the document nodes of all XML documents, residing in the /a/b/c/d directory of the filesystem, that satisfy your criteria (yourBooleanExpression evaluates to true())
An HTTP GET query string is a ordered sequence of key/value pairs:
?spam=eggs&spam=ham&foo=bar
Is, with certain semantics, equivalent to the following dictionary:
{'spam': ['eggs', 'ham'], 'foo': bar}
This happens to work well for boolean properties of that page being requested:
?expand=1&expand=2&highlight=7&highlight=9
{'expand': [1, 2], 'highlight': [7, 9]}
If you want to stop expanding the element with id 2, just pop it out of the expand value and urlencode the query string again. However, if you've got a more modal property (with 3+ choices), you really want to represent a structure like so:
{'highlight_mode': {7: 'blue', 9: 'yellow'}}
Where the values to the corresponding id keys are part of a known enumeration. What's the best way to encode this into a query string? I'm thinking of using a sequence of two-tuples like so:
?highlight_mode=(7,blue)&highlight_mode=(9,yellow)
Edit: It would also be nice to know any names that associate with the conventions. I know that there may not be any, but it's nice to be able to talk about something specific using a name instead of examples. Thanks!
The usual way is to do it like this:
highlight_mode[7]=blue&highlight_mode[9]=yellow
AFAIR, quite a few server-side languages actually support this out of the box and will produce a nice dictionary for these values.
I've also seen people JSON-encode the nested dictionary, then further encode it with BASE64 (or something similar), then pass the whole resulting mess as a single query string parameter.
Pretty ugly.
On the other hand, if you can get away with using POST, JSON is a really good way to pass this kind of information back and forth.
In many Web frameworks it's encoded differently from what you say.
{'foo': [1], 'bar': [2, 3], 'fred': 4}
would be:
?foo[]=1&bar[]=2&bar[]=3&fred=4
The reason array answers should be different from plain answers is so the decoding layer can automatically tell the less common foo case (array which just happens to have a single element) from extremely common fred case (single element).
This notation can be extrapolated to:
?highlight_mode[7]=blue&highlight_mode[9]=yellow
when you have a hash, not just an array.
I think this is pretty much what Rails and most frameworks which copy from Rails do.
Empty arrays, empty hashes, and lack of scalar value look identical in this encoding, but there's not much you can do about it.
This [] seems to be causing just a few flamewars. Some view it as unnecessary, because the browser, transport layer, and query string encoder don't care. The only thing that cares is automatic query string decoder. I support the Rails way of using []. The alternative would be having separate methods for extracting a scalar and extracting an array from querystring, as there's no automatic way to tell when program wants [1] when it wants 4.
This piece of code works for me with Python Backend-
import json, base64
param={
"key1":"val1",
"key2":[
{"lk1":"https://www.foore.in", "lk2":"https://www.foore.in/?q=1"},
{"lk1":"https://www.foore.in", "lk2":"https://www.foore.in/?q=1"}
]
}
encoded_param=base64.urlsafe_b64encode(json.dumps(param).encode())
encoded_param_ready=str(encoded_param)[2:-1]
#eyJrZXkxIjogInZhbDEiLCAia2V5MiI6IFt7ImxrMSI6ICJodHRwczovL3d3dy5mb29yZS5pbiIsICJsazIiOiAiaHR0cHM6Ly93d3cuZm9vcmUuaW4vP3E9MSJ9LCB7ImxrMSI6ICJodHRwczovL3d3dy5mb29yZS5pbiIsICJsazIiOiAiaHR0cHM6Ly93d3cuZm9vcmUuaW4vP3E9MSJ9XX0=
#In JS
var decoded_params = atob(decodeURI(encoded_param_ready));