- AppVeyor
- Travis CI
Now supports .Net Core, run same .Net code in Windows and Linux
- .Net Standard 2.0 > AdysTech.InfluxDB.Client.Net.Core (this is the preferred version)
- .Net 4.5.1 > AdysTech.InfluxDB.Client.Net
InfluxDB is new awesome open source time series database. But there is no official .Net client model for it. This is a feature rich .Net client for InfluxDB. All methods are exposed as Async methods, so that the calling thread will not be blocked. It currently supports
- Connecting using credentials
- Querying all existing databases
- Creating new database
- Querying for the whole DB structure (hierarchical structure of all measurements, and fields)
- Writing single, multiple values to db
- Retention policy management
- Post data to specific retention policies
- Query for all Continuous Queries
- Create a new Continuous Query
- Drop continuous queries
- Chunked queries
To be added are
- Query for all tags, their unique values
- Deleting data, currently drop queries can be fired to delete data
InfluxDBClient client = new InfluxDBClient (influxUrl, dbUName, dbpwd);List<String> dbNames = await client.GetInfluxDBNamesAsync ();Dictionary<string, List<String>> = await client.GetInfluxDBStructureAsync("<db name>");This returns a hierarchical structure of all measurements, and fields (but not tags)!
bool success = await client.CreateDatabaseAsync("<db name>");Creates a new database in InfluxDB. Does not raise exceptions if the DB already exists.
It represents a single Point (collection of fields) in a series. In InfluxDB each point is uniquely identified by its series and timestamps.
Currently this class (as well as InfluxDB as of 0.9) supports Boolean, String, Integer and Double (additionally decimal is supported in client, which gets stored as a double in InfluxDB) types for field values.
Multiple fields of same type are supported, as well as tags.
var valMixed = new InfluxDatapoint<InfluxValueField>();
valMixed.UtcTimestamp = DateTime.UtcNow;
valMixed.Tags.Add("TestDate", time.ToShortDateString());
valMixed.Tags.Add("TestTime", time.ToShortTimeString());
valMixed.Fields.Add("Doublefield", new InfluxValueField(rand.NextDouble()));
valMixed.Fields.Add("Stringfield", new InfluxValueField(DataGen.RandomString()));
valMixed.Fields.Add("Boolfield", new InfluxValueField(true));
valMixed.Fields.Add("Int Field", new InfluxValueField(rand.Next()));
valMixed.MeasurementName = measurementName;
valMixed.Precision = TimePrecision.Seconds;
valMixed.Retention = new InfluxRetentionPolicy() { Name = "Test2" };
var r = await client.PostPointAsync(dbName, valMixed);A collection of points can be posted using await client.PostPointsAsync (dbName, points), where points can be collection of different types of InfluxDatapoint
var r = await client.QueryMultiSeriesAsync ("_internal", "select * from runtime limit 10");
var s = await client.QueryMultiSeriesAsync("_internal", "SHOW STATS");
var rc = await client.QueryMultiSeriesAsync ("_internal", "select * from runtime limit 100",10);QueryMultiSeriesAsync method returns List<InfluxSeries>, InfluxSeries is a custom object which will have a series name, set of tags (e.g. columns you used in group by clause. For the actual values, it will use dynamic object(ExpandoObject to be exact). The example #1 above will result in a single entry list, and the result can be used like r.FirstOrDefault()?.Entries[0].time. This also opens up a way to have an update mechanism as you can now query for data, change some values/tags etc, and write back. Since Influx uses combination of timestamp, tags as primary key, if you don't change tags, the values will be overwritten.
Second example above will provide multiple series objects, and allows to get data like r.FirstOrDefault(x=>x.SeriesName=="queryExecutor").Entries[0].QueryDurationNs.
The last example above makes InfluxDB to split the selected points (100 limited by limit clause) to multiple series, each having 10 points as given by chunk size.
This library uses a cutsom .Net object to represent the Influx Retention Policy. The Duration concept is nicely wraped in TimeSpan, so it can be easily manipulated using .Net code. It also supports ShardDuration concept introduced in recent versions of InfluxDB.
var policies = await client.GetRetentionPoliciesAsync (dbName);The code below will create a new retention policy with a retention period of 6 hours, and write a point to that policy.
var rp = new InfluxRetentionPolicy () { Name = "Test2", DBName = dbName, Duration = TimeSpan.FromHours (6), IsDefault = false };
if (!await client.CreateRetentionPolicyAsync (rp))
throw new InvalidOperationException ("Unable to create Retention Policy");
valMixed.MeasurementName = measurementName;
valMixed.Precision = TimePrecision.Seconds;
valMixed.Retention = new InfluxRetentionPolicy () { Name = "Test2"}; //or you can just assign this to rp
var r = await client.PostPointAsync (dbName, valMixed);Similar to retention policy the library also creates a custom object to represent the Continuous Queries. Because the queries can be very complex, the actual query part of the CQ is exposed as just a string. But the timing part of CQ specifically given in [RESAMPLE [EVERY <interval>] [FOR <interval>]] are exposed as TimeSpan objects. Since the GROUP BY time(<interval>) is mandated for CQs this interval again is exposed as TimeSpan. This allows you to run LINQ code on collection of CQs.
var cqList = await client.GetContinuousQueriesAsync ();var p = new InfluxContinuousQuery () { Name = "TestCQ1",
DBName = dbName,
Query = "select mean(Intfield) as Intfield into cqMeasurement from testMeasurement group by time(1h),*",
ResampleDuration = TimeSpan.FromHours (2),
ResampleFrequency = TimeSpan.FromHours (0.5) };
var r = await client.CreateContinuousQueryAsync (p);var r = await client.DropContinuousQueryAsync (p);p has to be an existing CQ, which is already saved.