Dgraph: Data explorer for dgraph

[aksaha2311]

Currently I am missing a robust data explorer for dgraph. Like in agangodb, it is very easy to explore what data is there inside db. Which is must have feature in a DB.

[manishrjain]

How about Ratel? The binary is part of Dgraph releases.

[aksaha2311]

In my opinion, Retel is an advanced alternative to the terminal.

Only schema part is working as an explorer, rest we need to query to get the data.

There is no UI, which explore the data without any query. This kind of data explorer is very helpful to debug and validate our code. One of the examples of data explorer of MySQL is 'phpMyAdmin'.

[manishrjain]

CC: @paulftw .

Can you describe what exactly you'd like to see when you mean explore data without query?

[aksaha2311]

Basically I want to see all the data using GUI.

For example, I used phpMyAdmin as a data explorer for MySQL.

When I open phpMyAdmin, I can view all the databases I have in MySQL, then I can see all the tables the databases without using any SQL query, then I can see the data of the tables from there again without using any SQL query.

This kind of explorer will very much helpful at development time, where you have to see your query is doing what you expected.

[sboorlagadda]

I understand the need for an interactive UI to explore data, we can certainly start showing few nodes and provide a clickable option (+) on each node to explore the graph in a particular direction interactively (consider ratel automatically builds and runs a query when user clicks + on a node).

But it would be challenging for a single host (client browser or graph explorer) to handle huge graphs. In SQL data explorers (like phpMyAdmin) we can page through the results and not have to bring in all or huge data at once. But certain graph node can have a million edges, so this probably could be a reason to query a small subset.

[aksaha2311]

I have different opinion

certain graph node can have a million edges

At the time of development, developer use to test with small dataset, maybe at most 500 edges for a node. And if we can visualize this, most of the issues related to data or graph we can detect very easily.

[CreatCodeBuild]

The fact that Graphs are large isn't a good reason for not doing this visualization. It is certainly a challenge. But it will increase the adoption of Dgraph, which is critial for Dgraph's survival.

From a development perspective, tooling is 1st important reason for me to adopt a technology.

[AlexDunmow]

This would be a great tool to have, but I don't think it's necessarily the responsibility of the DGraph team to build it. Anyone with some decent frontend and Go skills could easily knock something up.

[shaneakr]

We’re going to start releasing prototypes and mocks on the features that are important to you on discuss. The first feature is a data explorer.
See mocks on discuss.dgraph.io

This mock is definitely an MVP and we already have a lot more ideas for functionality that we want to add but we want to get something out quickly that works.

We’d love to get your feedback.

[artemnl]

You can check the UI for the Neo4j. Even if there are a million nodes, it shows only 300 and you could open more if you REALLY want it.

[manishrjain]

Ratel now supports exploring data with the schema view, without issuing any queries. I think that's a sufficient fix for this issue. Feel free to open a new issue, if you want something specific in Ratel.