Manage data views
Shape a data object's grid with columns, filters, sorting, aggregations and formulas, and save it all as a view.
A view is a saved way of looking at a data object: which columns show, in what order, how the records are filtered and sorted, what each column totals, any formula columns you've added, and whether each row's label stays in sight as you scroll sideways. Every data object starts with a default view, and you can build as many more as you need. A view takes you straight to the slice of an object your work needs. Once saved, it is the same for everyone, so a team works from one cut of the data rather than each person filtering it a little differently. This page walks through each control on the grid and how to save your work as a view.
The grid's header
Open the sidebar's Ecosystem tab (⌘⇧E), click Data (click More first if the list ends at Canvas), and open a data object to see its record grid. Everything sits in one header band above the grid:
- On the left, the object selector and the view selector: switch to another object or another view of this one.
- On the right, the view's controls as icons, each named by its tooltip: Filter, Sort, Display, Formula and Merge. Then the pencil (Edit records), the ⋯ (More) menu, and + Record.
Under the header, a line shows the view's filters and sort whenever it has any, and below that the count of matching records (for example 2,068 results). At the right of the count, Clear view and Save view appear while the view has changes you have not saved, and Discard changes and Save changes while you have edited cells.
On a narrow window the right-hand controls move under the selectors, and + Record gives up its label for its icon.
The way you look at the records works the same whether or not edit mode is on: the pencil gates changes to the records themselves. Selecting records for bulk actions is part of edit mode (see Act on records in bulk), and so is editing cells; see Add, edit & import records.
What the view applies is always visible
A glance at the icons tells you whether the grid you are reading is narrowed, reordered or reshaped, without opening anything:
- Filter and Sort read as pressed while their line shows, which it does whenever the view filters or sorts. Each carries a small orange dot while the view applies something there, or has a change there you have not saved.
- Display reads as pressed, with its dot, while the view's columns differ from the default view's, and carries the dot while its columns have changes you have not saved.
- Formula reads as pressed and dotted while the view shows a formula column, and Merge while it shows columns from another object.
A view's filters and sort also show on their own line under the header as soon as it has any, so nobody mistakes a filtered view for the whole object.
Switch and manage views
The view selector sits at the top left, beside the object selector, and shows the current view's name. Click it to open the list.
- Switch view: pick any view from the list; the one on screen has a tick. Use the Search views... box to find one by name.
- Create a view: click Create new view at the bottom of the list, enter a View name, and click Create view.
- Duplicate a view: hover a view and open its ⋯ (or right-click the view), then choose Duplicate view. Name the copy (it suggests the view's name followed by (1)) and click Create view. The copy takes the view's columns, filters, sorting and aggregations, so you can branch off an existing setup. The default view can be duplicated even though it can't be renamed or deleted.
- Rename a view: the same ⋯ offers Rename for any view other than the default.
- Delete a view: the ⋯ menu's Delete view removes it after a confirmation. Deleting a view never touches the underlying records. If it was the view on screen, the grid goes back to the default view.
Choose which columns to display
Click the Display icon to manage columns. The menu opens on Manage view: Freeze reference display, Add column, then the columns the view shows, in order.
- Add columns: click Add column to open the attribute picker. Search with Search attributes..., tick the attributes to add, then click Apply at the foot, which counts what you picked. The columns the grid already shows are ticked; to take one out, use the list below.
- Remove a column: hover it in the Manage view list, open its ⋯ and choose Remove column, or open its header's menu in the grid and choose Hide from view. The same ⋯ has Rename, to give the column another heading in this view.
The default view lists the object's own attributes first, then the record's own columns, Created At, Created By, Updated At and Updated By, with the id column last, so a view never opens on a column of raw identifiers. The id is still there when you need it, at the far right.
Keep each row's label in view
When an object's records have a label (its Reference display), the grid pins that label at the left as you scroll sideways through a wide object, so you always know which row you are reading. Freeze reference display, at the top of the Display menu, turns this on or off for the view (a tick shows it is on), and it is saved with the view like its columns are.
Show a field from a related object
The column picker can reach across relationships, so a view can display data that lives on a linked record.
- In Add column, a relationship shows the object it points at and a chevron on the right. Click it to step into the related object.
- The picker's top line shows the path you followed: the object you came from, then the relationship (for example Fund › Fund Administrator). The back arrow returns you one level.
- Under As label, the relationship itself adds the linked record's label (its Reference display) as the column's value. Below it, tick any of the related object's attributes, then click Apply.
The new column is named after the path, for example Fund Administrator > LEI, so you can build a Funds view that shows each fund's administrator's LEI without leaving the grid. A column reached through a relationship is read-only: change the value on the linked record.
The column header menu
Each column header opens a menu with the actions for that column:
- Sort ascending, Sort descending and Disable sorting, the current one ticked.
- Calculate, to total the column (see Total a column).
- Move to left and Move to right, and Reset width on a column whose width you changed.
- Rename, to give the column a different heading in this view (the attribute itself keeps its name).
- Hide from view.
Reorder columns
Three ways to change column order, whichever feels natural:
- Drag a column header left or right in the grid.
- Drag a column's row in the Display › Manage view list; its grip shows it can be dragged.
- Use a column header's menu and choose Move to left or Move to right.
Filter
- Click the Filter icon. The filter line opens under the header, with a Filter button.
- Click Filter and pick the attribute to filter on (search with Search filterable attributes...).
- Choose a condition, then enter or pick a value.
Each filter on the line reads as one joined control: the attribute, the condition, the value, and × (Remove filter). Add another with + (Add filter); a record must match every filter on the line.
The conditions depend on the attribute's type:
| Type | Conditions |
|---|---|
| Text | contains, not contains, starts with, ends with, is, is not, empty, not empty |
| Number | equals, not equals, greater than, less than, greater than or equal, less than or equal |
| Date | is, is not, before, after, between |
| Boolean | is, is not |
| Relationship, user, option | includes, not includes, empty, not empty |
A date's value opens a calendar; between picks a range in it. The system Created At and Updated At columns filter as dates, rather than as text.
Boolean attributes are a value picker rather than a text box: choose Yes or No from the list instead of typing a word that has to match. Until you pick one, the value reads Value..., so a half-built condition is obvious rather than silently matching nothing.
A relationship opens on the related object, as in the column picker: pick one of its attributes to filter on it, for example a fund administrator's country, or pick the relationship under As label to filter on the linked records themselves.
Values are suggested as you type
A text or number filter suggests the values the column actually holds, the most recently seen first, so you rarely need to remember an exact spelling. Start typing in Type to search... and the list narrows to matching values, searched on the server across the whole object, not only the rows on screen. Use the arrow keys and Enter to pick one, or keep typing and choose Use “…” to filter on exactly what you typed.
The list shows up to 20 values. When the column holds more, Showing first 20 sits under the list: keep typing to reach the rest.
Picking values in a filter
Filters on relationship, user and option attributes open a searchable picker rather than a text field. The values you have already chosen are listed first, above the rest, so a long list never hides your own selection and unticking something does not mean hunting for it again. A relationship's picker then lists the records the column points at most recently; the rest load as you scroll, and the search reaches any of them.
Advanced filters
For anything beyond a list of filters that must all match, switch to the advanced builder: click Filter (or +) on the filter line and turn on the Advanced switch at the foot of the attribute list. The builder opens at once. From then on the line shows a single Filter button with an advanced badge, so you can tell at a glance that the grid is scoped by more than the conditions on show; click it to open the builder again. The builder gives you full boolean logic:
- Each condition sits beside a logic label. The first row reads Where; the second picks And or Or for its group, and the rows after it repeat that choice.
- Add filter adds another condition; Add group starts a nested group with its own And or Or, so you can express things like (Domicile is Luxembourg And NAV greater than 500) Or Umbrella is No. Conditions can also reach across relationships, for example Fund Administrator > Country is Luxembourg.
- Each condition's ⋯ has Remove. Remove a nested group with Delete group, and reset everything with Clear all filters.
Turn the Advanced switch off to go back to the filter line; the builder's conditions don't carry over to it.
An advanced filter counts everywhere a filter does: the export writes only the records it matches, and changing it offers to save the view, just like a simple filter.
Filtering a very large object on a column SPARK does not index for itself (a valuation date, say) can be slow, because every read scans the whole table. If a filter you use daily is slow on a large object, ask NGT support to index that column: an index can be built on one named column on demand, without stopping writes.
Sort
- Click the Sort icon. A Sort button appears on the line under the header; click it to open the sort editor.
- Click Select an attribute..., choose an attribute, and pick Ascending or Descending.
- Click Add sort to add another level.
- Drag a level by its grip to change which sort applies first; the bin (Remove sort) takes a level out.
The sort applies when the editor closes. The line's button then names it: Sorted by and the leading attribute, with a count of the levels after it (for example Sorted by Domicile +1), so the active order is always visible, and the Sort icon carries its dot.
When two attributes share a display name, the attribute list shows each one's slug beside it so you can tell them apart. Attributes with a name of their own show no slug, since they need no disambiguating.
Large objects load as you scroll
The grid reads records from the server a page at a time as you scroll, rather than loading the whole object, so an object with millions of rows opens as quickly as a small one. Filters and sorts run on the server too. The first rows show before the total is counted; until the count arrives, a placeholder sits where it goes.
Total a column with an aggregation
Any column can show an aggregate in the totals row pinned at the bottom of the grid.
- Open a column header's menu and choose Calculate.
- Pick an aggregation. Number and formula columns offer Count Filled, Count Empty, Count Distinct, Sum, Average, Min and Max; other columns offer the three counts. The record's own columns (id, Created At and the like) have no Calculate.
- The result appears in the bottom totals row, with a short word naming the metric (sum, average, filled and so on). Choose None to clear it.
The total is worked out on the server, across every matching record rather than only the rows loaded on screen. Aggregations are part of the view, so save the view to keep them.
Show columns from another object
The Merge icon borrows columns from a second data object and shows them beside your own, so a view can answer a question that spans two objects without exporting anything. Your row count never changes and the borrowed columns are read-only.
See Merge with other data for the full walkthrough.
Add a formula column
The Formula icon adds a computed column to the current view without changing the object's stored attributes.
- Click the Formula icon.
- Enter a Column name.
- Write the expression under Formula. Column names autocomplete as you type, and the hint shows the shape, for example
SUM(col_1, col_2). The ? beside Formula lists the functions:SUM,MULTIPLY,DIVIDE,SUBTRACT,AVERAGE,ABS,POWER,CONCAT,UPPER,LOWER,SUMLIST,AVGLISTandIF.SUMLISTandAVGLISTsum and average a list field, such as a value on each related record. - Click Apply to add the column.
The new column is computed for every row and is read-only. It lands before the record's own columns (Created At and the like), and it lives in the view, so save the view to keep it. If you instead want a permanent formula stored on the object itself, create a Formula attribute in the object's settings; see Attribute types.
Act on records in bulk
Selecting records is part of edit mode. Click the pencil (Edit records) first, and checkboxes appear at the left of the grid. The header checkbox selects everything loaded, and if more rows match your filters a prompt offers Select all N matching your filters. While records are selected, a bar takes the place of the count with N selected and the actions on the right:
- Edit records sets one attribute to the same value on every selected record. See Edit one field on many records.
- Archive hides the records from active views without deleting them.
- Restore brings archived records back, when the selection includes any.
- Delete removes them permanently, after a dialog that previews the impact (including how many related records are detached or also deleted). See Add, edit & import records for the detail.
× (Clear selection), at the bar's left, drops the selection.
See and restore archived records
Archived records are hidden by default. To work with them, open the ⋯ (More) menu and tick Show archived: the archived records join the others in the grid, in the same order, greyed and read-only. In edit mode, select the ones you want and click Restore to bring them back into active views, then untick Show archived to return to the normal grid. Show archived does not reach a view filtered with the advanced builder.
Export the view as a CSV
Open the ⋯ (More) menu and choose Export data, then click Export. The file downloads as a CSV named with the date and time of the export, for example export_20260812T093000.csv.
Export uses the current view, so the file matches what's on screen: the active filters (simple or advanced) decide which records are included, archived ones too while Show archived is on; the sort sets the row order; and the visible columns, in their order on screen, decide which attributes are written out. A relationship column is written as the linked records' names, not their ids. Adjust those first to shape the file.
Export writes the first 10,000 records that match. The dialog reads Export the first 10K records to a file, or, when more match, Export the first 10,000 of N matching records to a file; after the download, a message says only the first 10,000 were exported. Narrow it with filters and export in batches.
Save your work as a view
Changing columns, filters, sorting, aggregations, formulas or the frozen label marks the view as edited, and two buttons appear at the right of the count:
- Clear view drops your edits and returns to the view as last saved.
- Save view updates the current view in place, for everyone. Its arrow (More save options) offers Save to this view and Save as a new view, which keeps the current view as it was and stores your changes under a new name.
The default view can't be overwritten: on it, Save view asks you to name a new view instead, and Save to this view is greyed out with the reason. In the Create new view dialog, enter a View name (it suggests the view's name followed by (1)) and click Create view.