Skip to main content

Merging Data Sources

About 3 mindataworkspace

Merging Data Sources

IOTA can display data from multiple sources in the same view — PI historian data alongside Seeq calculations, SQL database records next to real-time sensor values, or signals from two different sites on the same trend. Each component channel has its own data source selector, so mixing sources is a per-channel decision.

When You Need This

Consider a multi-source display when:

  • An overview display needs KPI values from your historian alongside contextual metadata from a SQL database (batch records, production targets, shift schedules)
  • You want to show live process data from PI next to advanced analytics or lab results from Seeq
  • Two sites use separate PI servers and you need side-by-side comparison in one trend
  • A calculation exists in Seeq that is not available as a raw PI tag

How It Works

Every component in IOTA has one or more channels. Each channel represents one data stream shown by that component. Each channel independently specifies:

  • Which data source to query
  • The tag, asset attribute, or dataset to pull from that source

A single Line Chart can have four channels — two pulling from a PI server, one from Seeq, and one from Azure Data Explorer — and they all render on the same chart with the same time axis. The GTC applies to all time-aware channels regardless of source.

Setting the Data Source Per Channel

  1. Select a component and open the Settings panel.
  2. Navigate to the Channels tab.
  3. For each channel, use the Data Source dropdown to select the source.
  4. Open the Search panel, switch to the matching data source, then search and drag the tag, attribute, or dataset onto the channel.

A Line Chart fed by channels from two sources — PI_PROD and SQL_BATCH — on one trend
A Line Chart fed by channels from two sources — PI_PROD and SQL_BATCH — on one trend

Repeat for each channel that needs a different source. The component renders data from all channels together once each source returns its result.

What to Expect by Source Type

Response time and behavior vary significantly across source types. Build with these differences in mind:

Source typeBest forTypical latency
PI / AF HistorianReal-time process data, high-resolution trendsSub-second
SeeqAdvanced calculations, lab context, signal analysisSeconds
SQL / Relational DBStructured records, batch data, metadataSeconds to tens of seconds
DatabricksAnalytics workloads, large-scale batch and ML dataSeconds to minutes
Azure Data Explorer / SnowflakeLarge-scale historical queries, analytics workloadsSeconds to minutes

Tips

For real-time operational displays, use PI or another historian as the primary data source. Pull Seeq and SQL data for context or supplementary analysis, not as the main signal feed. Limit time windows on SQL and analytics sources to keep load times acceptable — a wide time range against a slow source can stall the entire display render.

Planning Your Data Source Configuration

Before building a multi-source display, confirm with your administrator which sources are active in your IOTA instance and which source name maps to which system. Use a simple table to document this:

Source name in IOTASystemCloud / On-PremWhat it contains
PI_PRODOSIsoft PI ServerOn-PremReal-time process historian
SEEQ_MAINSeeq CloudCloudCalculations, lab data
SQL_BATCHSQL ServerOn-PremBatch records and production targets

Source names are set by the administrator. They may not match the underlying system name exactly. Getting this mapping from your admin before building saves time.

A source only appears in IOTA after its driver has been configured. For the full list of supported connectors and their setup, see Data Source Drivers.

Go to Seeq

For any component with at least one channel bound to a Seeq-connected data source, you can open the current signals and time range directly in Seeq Workbench. Right-click the component while in locked (viewer) mode and select Go to Seeq. See Go to Seeq for supported components and requirements.

When a Display Loads Slowly

A multi-source display renders at the speed of its slowest source — one slow historian query or analytics scan holds up the whole render. To find which source is responsible, check the Driver Log and diagnose by source type in Troubleshooting → Diagnosing Slow Loads.

What's Next

With your data sources merged into one display, the next step is composing more complex layouts — embedding other views and organizing content into tabs. Continue to Organizing Complex Displays.

Driver documentation (administrator setup required before a source appears in IOTA):

IOTASoftware Licensed | Copyright © 2021-present