NIFI-4989 Made PutMongo able to use nested lookup keys, a query param… by MikeThomsen · Pull Request #2560 · apache/nifi

MikeThomsen · 2018-03-17T01:53:27Z

… and multiple lookup keys.

Thank you for submitting a contribution to Apache NiFi.

In order to streamline the review of the contribution we ask you
to ensure the following steps have been taken:

For all changes:

Is there a JIRA ticket associated with this PR? Is it referenced
in the commit message?
Does your PR title start with NIFI-XXXX where XXXX is the JIRA number you are trying to resolve? Pay particular attention to the hyphen "-" character.
Has your PR been rebased against the latest commit within the target branch (typically master)?
Is your initial contribution a single, squashed commit?

For code changes:

Have you ensured that the full suite of tests is executed via mvn -Pcontrib-check clean install at the root nifi folder?
Have you written or updated unit tests to verify your changes?
If adding new dependencies to the code, are these dependencies licensed in a way that is compatible for inclusion under ASF 2.0?
If applicable, have you updated the LICENSE file, including the main LICENSE file under nifi-assembly?
If applicable, have you updated the NOTICE file, including the main NOTICE file found under nifi-assembly?
If adding new Properties, have you added .displayName in addition to .name (programmatic access) for each of the new properties?

For documentation related changes:

Have you ensured that format looks appropriate for the output in which it is rendered?

Note:

Please ensure that once the PR is submitted, you check travis-ci for build issues and submit an update to your PR as soon as possible.

… and multiple lookup keys.

zenfenan

Requesting changes only for the description to be added in the new UPDATE_QUERY PropertyDescriptor.

zenfenan · 2018-03-18T06:06:50Z

        .defaultValue("_id")
        .build();
+    static final PropertyDescriptor UPDATE_QUERY = new PropertyDescriptor.Builder()
+        .name("putmongo-update-query")


How about adding description here?

zenfenan · 2018-03-18T06:13:02Z

+                    query = parseUpdateKey(updateKey, (Map)doc);
+                    removeUpdateKeys(updateKey, (Map)doc);
+                } else {
+                    query = Document.parse(updateQuery);


Shall we change it to filterQuery ? Purely a cosmetic change, feel free to ignore it but I think it follows a better naming convention since it's actually a filter query based on which the update is done.

zenfenan · 2018-03-18T06:14:27Z

+        updateTests(document);
+    }
+
+    @Test


If possible, would love to see a similar test but with PutMongo.UPDATE_WITH_DOC

Yeah. I'm adding 2 at the moment. One for update by keys and one for query.

pvillard31 · 2018-03-18T11:02:04Z

-        .required(true)
-        .addValidator(StandardValidators.NON_EMPTY_VALIDATOR)
+        .required(false)
+        .addValidator(Validator.VALID)


Since required is set to false, you can still use the original validor: it'll be used only if a value is set. No? Unless you really want someone to be able to set an empty string as a possible value.

Wasn't sure if that's the case.

pvillard31 · 2018-03-18T11:05:34Z

+        List<ValidationResult> problems = new ArrayList<>();
+
+        final boolean queryKey = validationContext.getProperty(UPDATE_QUERY_KEY).isSet()
+                && !StringUtils.isBlank(validationContext.getProperty(UPDATE_QUERY_KEY).getValue());


You don't need to check that if my previous comment is valid (since, if a value is set, the validator will check that's the value is not empty)

I just remembered why I did this. When you call removeProperty to unset the property, the getProperty call here will return the default value. So maybe I need to remove the default value. What do you think? Is this a problem with the test helpers or my understanding of how properties should work?

zenfenan · 2018-03-18T13:35:57Z

+    private void removeUpdateKeys(String updateKeyParam, Map doc) {
+        String[] parts = updateKeyParam.split(",[\\s]*");
+        for (String part : parts) {
+            if (part.contains(".")) {


Forgot to raise this earlier. Why are we doing a check if it contains a . ? It will work only for complex keys and simple key will be ignored i.e. not removed, right?

For this input:

{ "name": "John Smith", "department": "Engineering" }

It makes no sense to remove name if we're doing a full document update using the `name key.

Now consider this complex document:

{ "name": "John Smith", "department": "Engineering", "contacts": { "email": "john.smith@test.com" } }

To search on email, we have to submit this payload with the lookup key being contacts.email:

{ "contacts.email": "john.smith@test.com", "name": "John Smith", "department": "Engineering", "contacts": { "email": "john.smith@test.com" } }

Mongo cannot do a lookup using this: { "contacts": { "email": "john.smith@test.com" }}

So if we don't remove the complex lookup key, we are leaving extraneous information in the document that almost certainly has no value to the user.

Now maybe we'll get an angry ticket complaining that they can't do periods in the key names, but I've never seen normal use cases where developers do that. The whole idea of creating complex key names for real data using periods and such flies in the face of how JSON is supposed to work.

No.. what I meant was, regardless of whether we have a complex key or simple key, we don't need them to exist in the parsed filter query, right? So why don't we simply remove this if (part.contains(".")) check, basically meaning, remove the key from the doc. No?

we don't need them to exist in the parsed filter query, right?

We do when replacing one document with another. Say you have a user account document, and the update key is username. If the user is doing full document replacement, and we remove the update key, the username field is going to be removed from the document that is used in the substitution. So then the poor end users will be searching through an accounts collection looking for documents with no usernames.

Also, again, complex key names simply wouldn't have any value to over 99% of Mongo users. As I said, I'm sure there's a few random users that want to chuck something like this in Mongo:

{ "user.name": "john.smith", "user.contact.email": "john.smith@test.com" }

Accommodating that would be detrimental to providing a consistent tool for the vast majority of Mongo users.

Understood. So this is critical when doing replace i.e. the mode is set to UPDATE_WITH_DOC. Correct?

Yes. The only time it wouldn't apply is if a user wanted to remove a field like they misnamed the username field as yusername. They can fix that with:

{ "yusername": "....", "$set": { "username": ".....", }, "$unset": { "yusername": 1 } }

TBH, the sanest way to handle such mistakes would be to stop ingestion and run some JavaScript code to update the database properly.

MikeThomsen · 2018-03-20T12:14:34Z

@pvillard31 Thoughts on merging this?

pvillard31 · 2018-03-20T13:09:11Z

@MikeThomsen - can you have a look at this commit: pvillard31@5a67aa1

I reworked a bit the tests to avoid "removing" a property while executing the tests. Basically, instead of having a shared instance of the TestRunner for all the test suite, I initialize one each time (to be sure there is no conflict between the tests and to avoid issues in cases tests are not executed in the same order each time). Regarding your comment about the validation failing, it's because we're providing a default value "_id". Because of that the property will always return true to isSet(). Since we're adding a new property and the user needs to use one of the two, I suggest removing the default value: it does not break anything, and it "forces" the user to understand the choice the user is doing. Does that make sense?

MikeThomsen · 2018-03-20T14:21:51Z

I think that works.

zenfenan

Looks good to me. Thanks @MikeThomsen

MikeThomsen · 2018-03-23T11:56:31Z

@pvillard31 Can you merge your change so we can close this out?

pvillard31 · 2018-03-23T13:47:31Z

Done, merged to master, thanks @MikeThomsen @zenfenan

NIFI-4989 Made PutMongo able to use nested lookup keys, a query param…

380e342

… and multiple lookup keys.

zenfenan suggested changes Mar 18, 2018

View reviewed changes

pvillard31 reviewed Mar 18, 2018

View reviewed changes

NIFI-4989 Added some additional tests and fixes based on a code review.

a924104

zenfenan reviewed Mar 18, 2018

View reviewed changes

zenfenan approved these changes Mar 20, 2018

View reviewed changes

asfgit closed this in 66590b7 Mar 23, 2018

MikeThomsen deleted the NIFI-4989 branch August 14, 2024 21:14

Conversation

MikeThomsen commented Mar 17, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

For all changes:

For code changes:

For documentation related changes:

Note:

Uh oh!

zenfenan left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

MikeThomsen commented Mar 20, 2018

Uh oh!

pvillard31 commented Mar 20, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

MikeThomsen commented Mar 20, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

zenfenan left a comment

Choose a reason for hiding this comment

Uh oh!

MikeThomsen commented Mar 23, 2018

Uh oh!

pvillard31 commented Mar 23, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

MikeThomsen commented Mar 17, 2018 •

edited

Loading

pvillard31 commented Mar 20, 2018 •

edited

Loading

MikeThomsen commented Mar 20, 2018 •

edited

Loading