* add Locationa and Type Stringer

* Test errors as json, not require.Equal

* All queries w...

This fix builds on top of the fixes done in
bf1e288bfffa2510b1c20f1353b85f40eb322f49.

Replac...

* Change x.Parse to return a struct instead of a pointer.
* fix tests.

Previously, the JSON chunker was generating the same blank node for different map objects. This ...

Update tests that use pagination to also use sorting so the results are
deterministic based on ...

- Remove query cache which is causing contention.
- Make geo filtering execute in parallel.
- ...

This adds --ratel and --ratel_port flags to run Ratel via the compose tool.

Run default compo...

The current JSON chunker/parser was creating a slice of NQuads for each JSON map. This resulted ...

* Add request id into context and use in logs and errors

I had an issue since GOBIN and GOPATH were both set in my environment to
simply setting GOPATH ...

This change includes the [breaking/format] change dgraph-io/badger@88d5a3c that changes the unde...

Added CHANGELOG updates for v1.1 up to commit 5ea255340.

- Add stat for how long it takes to retrieve to get a timestamp from
zero.
- Simplify encoding...

* Adding DGRAPH_ENDPOINT env variable

* Using relative wiki path

* Pick DGRAPH_ENDPOINT fr...

* Bump Go version for release builds to 1.12.7.

* Run Ratel tests.

These tests used to run...

Use DestUIDs() function while accessing uids from child in GroupBy query. This fixes #3768.

This takes advantage of TeamCity's Go test integration, such as grouping tests per package.

h...

* Let's reuse postings
* Do not use fmt.Sprintf to generate collision key. Instead use and stor...

* Add Type to Field interface

* Add value completion algorithm to resolver

* Use value com...

* Adding restretto/z in vendor

* Adding code for restretto

* Remove schema equality functions

Remove all the equality check functions. Sort keys of sche...

We used to create a new lexer instance for parsing every RDF line in bulk loader. This used to r...

* Use Go subtests for 21-million data set queries so that each query
test can be captured by ...

Use DestUids while doing comparison for filters

This allows filtering subgraph using the length of a uid list stored in uid variables. This chan...

This PR has a significant impact on the allocated objects and space. For addMapEntry it went dow...

- Fixes the race conditions mentioned in #3685 by removing global variables from `worker/task.go...

In #3708 (4b41d9c907f3cef4ab83bbbdc1f9ba86f354134d) we bumped up the Raft
heartbeat rate to 100...

Change supported scalar to struct. Wrap schema generation internals
to a single function. Add t...

* Fixup how response is written and add tests

* Create dgraph package
- move dgo client cod...

This fixes the bug where we were exiting out of rebalanceTablets goroutine after the first run i...

Ristretto will be using the z package from now on so the z package in
dgrpah (containing utilit...

- Fix config file to exclude issues like "type Uid should be UID".
- Remove unused warnings for...

Use the $QUERY_DIR absolute path instead of the relative path for the queries
directory so that...

This test is currently creating its own cluster. This PR modifies the
test so that it uses a cl...

To further #3695 , only read posting lists in certain cases. For e.g., when they are indexed, or...

From and to arguments for a shortest path query could only take hard coded uid values earlier. T...

The current logic only adds the '<' and '>' characters around the
predicate list if it has some...

Adding index mutations can be performed without reading the posting
lists in disk. This change ...

The filesystem backup tests are a bit flaky in TeamCity. As a first
step, this change adds veri...

./test.sh -C or ./test.sh -F can be used to run only the custom
cluster tests or only the long ...

Before we used to have 20ms ticks, but they would overload the
Raft tick channel, causing "A ti...

This reverts commit a1e36819367577cf4b898d01163d3fea3a54f113.

This reverts commit aeb892574a693298c572081befe9d73bfe9cfea4.

User specifies a schema having input types. The schema generator takes
that as input and create...

* Removing the first empty line and comma in the json output

* Moved the separator logic into...

Not sure if this was causing issues before, but rebuilding the count
reverse index depends on t...

The stream framework can iterate over the index prefix keys concurrently
whereas currently ther...

Currently, the time taken to receive the query response is included as
part of the encoding tim...

Dgraph is able to store any kind of valid geodata, but will only index
Point, Polygon, and Mult...

Specifically, the default value of depth when loop is set to false was
not documented.

If more than one uid is found in the list for a non-list predicate,
print a warning and force t...

The Fast Data Loading scetion used to be called Bulk Data Loading. It was
renamed in c871f604095...

Bump up the number of logs per span to capture all the events
from Server.Query which can creat...

* Add --tag flag to set dgraph/dgraph image tag. It must be used
with `-l=false` to use the i...

The assumption is not actually true since a DropAll can happen after the
first index check.

This was previously undocumented and I had to dig into the code to
confirm that certain JSON ob...

For some reason, there are nil values stored in the uidMatrix, which
results in a segmentation ...

Bug fix for bulk loader changes introduced in #3542.

Fixes #3607.

Signed-off-by: பாலாஜி ஜி...

The crash happens because querying this predicate does not populate the
uid matrix, resulting i...