Let's say I have some data that I obtained through a non-graphql endpoint for example from third party server (firebase).
How do I put the data into the local relay store?
Is there an easy way to add / edit / overwrite data to relay store directly without going through query or mutation?
A non public RelayStoreData field is accessible from the Relay.Store instance and it gives you direct access to the records contained in the store. I haven't done anything with this myself but you could try modifying the cache directly like this:
RelayStore._storeData._cachedStore._records[recordId][fieldName]=newValue
I would use relay without a server, defining your graphql schema locally and doing your API requests from your graphql schema the same way you would query a database in your schema.
https://github.com/relay-tools/relay-local-schema
Related
I have a task toget some data from an external supplier.
They have a Rest OData API. I have to connect using a subscription-key(APIKey).
When creating the OData LService, I add an Auth Header: "subscription-key" and in the Value field, I enter my key. After saving, I create a new dataset, and the OData LinkedService, provides me with the remote tables. I can choose the table I want and after that I create a pipeline to copy data from that table to my Azure SQL server.
This works fantastic :-)
However, after closing my browser and re-open it, the subscription key that I have entered earlier on the linked service, is now replaced with stars as it is a securestring. When I now run my pipeline, it will think that my key is the ten stars that have replaced my real key.
What am I doing wrong here ?
Also I would prefer to get my value from the KeyVault, but it seems that this is not possible on ODat connections....
Hope someone is able to provide some insight here :-)
BR Tom
From my testing I did not get any error on re-running. However coming to dynamic keys - I was not able to achieve it using the ODATA linked service.
Alternatively, if you can hit the ODATA endpoint with REST / HTTP Connector
You could - have a Web Activity to get the keys from the Key Vault and Set in the Variable.
WEB Activity URL : https://<your-keyvalut-name>.vault.azure.net/secrets/<your-secret-name>;
You could access the output of the web Activity using : #activity('Web1').output.value & Store in a variable.
You can reference this variable as the SUBSCRIPTION KEY for the subsequent steps in the REST/HTTP dataset.
You could pass it along the additional headers
I am replicating the Successfactors Employee Central Data (including FO, MDF, BG elements and etc.) via OData API to local database for third party integration.
It is able to trace changed records by filtering last modify date. However, the deleted record is not able to capture from OData API. Hence I cannot delete the record in my local database when corresponding EC record is deleted.
Is there any way I can get the deleted records from the API or other sources? Thanks.
OData API is not able to handle this task.
Extract from SF OData API doc:
Don't use our OData APIs when:
● Your system cannot consume either OData APIs or SOAP for an initial data load. In this case, you would go
for Import/Export with a CSV. Automation via FTP would also be a possibility.
● You need employee replication field level delta, snapshot, or read modified employees only, then SOAP Compound Employee API is your tool of choice. You can find more information in the guide Implementing the Employee Central Compound Employee API.
● You only need to read data, then the SOAP Compound Employee API would also be your tool of choice.
However with SFAPI (SuccessFactors CompoundEmployee API) it's easy. SFAPI has a special parameter changedSegmentsOnly that does exactly just what you want, the API returns only changed segments with an action code not equal to NO_CHANGE in delta transmission.
You make a query, for example, for changed employee data:
<urn:query>
<urn:queryString>select person,
personal_information,
address_information,
email_information,
phone_information,
employment_information,
job_information,
compensation_information,
paycompensation_recurring,
paycompensation_non_recurring,
direct_deposit,
national_id_card,
payment_information
from CompoundEmployee
where person_id_external = 'cgrant'
</urn:queryString>
<urn:param>
<urn:name>resultOptions</urn:name>
<urn:value>changedSegmentsOnly</urn:value>
</urn:param>
<urn:param>
<urn:name>maxRows</urn:name>
<urn:value>50</urn:value>
</urn:param>
</urn:query>
This API query will return you all the employees that were changed or deleted.
After that you can filter your response by that field.
ADDENDUM: any change to MDF entity, both standard and custom, can be tracked via OData Audit logs. How to enable them:
Go to this setting in API center
Enable this switch. It can be enabled for SFAPI as well
This way all API calls payloads will be saved and you will be able to see what entity was changed with each call
Prerequisite: the object must be visible by API and MDF version history must be enabled
More about API types for SuccessFactors:
SAP Note 2613670
OData API reference Guide
SFAPI reference Guide
I am looking for a way to export breeze entities on the server side to a json string which a breezejs manager can import from the client side. I looked all over the breeze APIs (both public and internal source code) but I couldn't find an obvious way of achieving this. There is a possibility of getting the desired results by using BreezeSharp (a .NET breeze client) on the server side but I would like to see if this is achievable with using the breeze server APIs only.
First you need to determine the shape of the bundle to be imported, i.e. something that manager.importEntities will understand. I don't think the format is documented, but you can reverse-engineer it by using:
var exported = manager.exportEntities(['Customer', 'Product'], {asString:true, includeMetadata:false});
Then pretty-print the value of exported to see the data format. See EntityManager.exportEntities for more info.
Once you have that, you can re-create it on the server. In C#, you can build it up using Dictionary and List objects, then serialize it using Json.NET.
An alternative approach would be to have your webhook just tell the client to initiate a query to retrieve the data from the server.
Let's say there I have an app that has model structure "Teams" and it's respective "members". This app then pulls data from a web service in json. How does the web service communicate which data has changed and how?
I'm thinking it passes certain JSON keys to notify the app which kinds of updates have occurred. For example:
{"operations": [{"delete":"member1"}, {"add": team2}, {"rename": team3} ... ]}
What are the conventions of doing this?
Edit 1: I am not looking for frameworks that solve this problem. I just want to conceptually know how this is usually done.
Generally the app will ask the server: "What has changed since XXX" and the server will reply with the objects that have changed since that date. The server generally gives full objects for the app to parse and consume.
With that assumption you can get the response from the server and then walk the objects in the JSON payload, loading objects that exist and updating them and then inserting new objects that do not exist locally.
Deletes can be more challenging as most servers that I have seen won't tell you about them. If you control the server then you could send a response with identifiers to objects that have been deleted since the last update.
I would like to create a communication API between my website which was built using rails and another website, so I can send data from my database to there website. So my idea is to use node.js where I will create real-time, so it will work in the following procedure. My Database is connected to rails, rails will send it to redis and then redis will send it to node finally node will send it to the other website using socket.IO
The question: Is that valid? Is there easier way?
Use just rails, create a controller and action that returns a json.
You can use https://github.com/rails-api/rails-api to speed things up.
I don't think that is a good way
You should consider lowering your dependency chain (DB -> Rails -> Redis -> node -> client)
Is redis necessary? (you can connect node to your DB, if you need realtime)
And if you don't need a realtime system, consider using http instead, you can return any type in rails (your own json, html, ...)