Skip to content
| Marketplace
Sign in
Visual Studio Code>Data Science>Data TablesNew to Visual Studio Code? Get it now.
Data Tables

Data Tables

PumasAI

| (0) | Free
Opens CSV, Parquet, Arrow, SAS, Stata, SPSS and Excel files, and JSON records, as tables that load only the rows in view.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

Data Tables

View data files as tables in VS Code, from SAS datasets to CSV and Parquet. Sort and filter the rows, and see a summary of every column beside them.

CDISC pilot demographics filtered to women aged 65 and over and sorted by age, with the summary panel open on the AGE column's profile

Supported files

These open as tables when you open them in VS Code:

Format Extensions
SAS .sas7bdat, .xpt
Stata .dta
SPSS .sav, .zsav, .por
Excel .xlsx
Delimited text .csv, .tsv, .tab, .psv, also compressed as .gz or .zst
Parquet .parquet
Arrow and Feather .arrow, .arrows, .feather, .ipc

To open another file as a table, right-click it and choose Open With...:

  • Data Table (read as delimited text) suits NONMEM output tables and other text files with columns.
  • Data Table (read as JSON records) reads an array of objects (.json) or one object per line (.jsonl, .ndjson).

Working with a table

Click a column header to sort by it, and Shift-click another header to add it to the sort. The toolbar spells out the sort, such as "Sorted by AGE ↓, SEX ↑". Click a column there to scroll the table to it, or the × to go back to the file's order. Hover over a header, or tab to it, to show its ⋯ menu button. A column's menu also pins the column or hides it. Click a row to select it. Shift-click selects a run of rows, and Cmd-click (Ctrl-click on Windows and Linux) adds one. Cmd+C or Ctrl+C copies them, ready to paste into a spreadsheet.

Filtering

The filter row under the toolbar holds your filters as chips. + Filter, or Filter... in a column's menu, opens a form where you pick a column and a condition and type the value. Click a chip to change it, or its × to remove it. The status line gives the rows shown and their share of the file, such as "179 of 306 rows (58.5%)".

The conditions fit the column's type. "Is one of" lists the column's values with their counts, and their labels for a coded column. Text matches ignore case unless you tick Match case.

Column Conditions
Any equals, does not equal, is one of, is not one of, is missing, is not missing
Numbers, dates, timestamps less than, at most, greater than, at least, between, not between
Text contains, does not contain, starts with, ends with, matches regex
True or false is true, is false

The filter form on the ARM column with "is one of" ticking the two Xanomeline arms in a list of values with their counts

A condition such as "does not equal" keeps the rows where the value is missing. Add a second chip, "is not missing", to drop them too. Untick Enabled in a chip's form to keep the chip without applying it. With two chips or more, Match all keeps the rows that meet every chip, and Match any keeps the rows that meet at least one. Clear all removes every chip.

SQL switches the row to a DuckDB SQL condition, such as AGE > 40 AND SEX = 'F'. The box starts with your chips written as SQL, and colours the SQL as you type, with the table's columns marked. Filters brings the chips back, and asks first if you changed the SQL.

SQL mode with a long coloured condition wrapping over two lines, and the 69 rows it keeps

The summary panel

The summary panel beside the table lists every column with a small chart of its values and its share of missing values. The chart is a histogram for numbers and dates, and the most common values for anything else. Its counts cover the rows your filters keep.

Click a column in the panel to see its full profile. The profile gives the missing and distinct counts and the top values. For numbers and dates it adds a histogram and summary statistics, such as the mean, the geometric mean and the quartiles. Profile in a column's menu opens the same view.

The summary panel with AGE's profile and histogram open, above other columns with their labels and small charts

Compare with, under a profile, plots the column against another one. Two columns of numbers or dates draw a heatmap. A column of text, or one switched to its top values, splits the other's histogram into one per value. Each of those histograms scales to its own rows, so their shapes compare. Two columns read by their values draw a grid of counts.

AGE compared with ARM: a histogram of ages for each arm, from Xanomeline Low Dose to Screen Failure

Search the panel by name, or sort it by name or type. Hovering a row swaps its type icon for an eye, which hides the column from the table or shows it again, and the → button scrolls the table to that column. A hidden column's row keeps its struck-through eye. While the panel is closed, the toolbar counts the hidden columns, and a click on the count opens the panel. The panel opens at most a third of the table's width. Dragging its edge resizes it as wide as you like, while the table keeps 240 px for its rows, and a narrower editor shrinks the panel to keep that room. The Toggle Summary Panel button in the editor's title bar shows or hides the panel.

SAS, Stata and SPSS labels

Files from SAS, Stata and SPSS keep their labels. A file's own label, such as "Demographics", shows in the toolbar before the row count. The summary panel shows each variable's label under its name, and hovering over a column's header shows it too. A coded column's menu offers Show value labels, which shows "Male" in place of 1, and Show codes to switch back. Sorting and filters still use the codes.

A special missing value keeps its code: a SAS .A, a Stata .a or an SPSS user-defined missing such as 98.0. The cell shows the code in grey, or its label when the column shows labels. Statistics count it as missing.

A Stata file showing value labels such as Male and Female, a grey "Not asked" for a missing code, and variable labels in the summary panel

A table opens again as you left it, with its sort, filters and columns. The summary panel stays open or closed as you left it, and a table left in SQL mode opens with its SQL.

Settings

Search for "Data Tables" in VS Code's settings. They set how cells look, how numbers and dates format, and how delimited text reads by default. Summary Position puts the summary panel on the right or the left, and Summary Open sets whether a table opens with it.

Your data

Data Tables never changes your files and never connects to the internet.

Platforms

macOS on Apple silicon, Windows x64, and Linux x64 and Arm64.

How it works

DuckDB reads each file in a process separate from VS Code, and the table fetches only the rows on screen. A file of millions of rows opens as fast as a small one. SQL mode takes a DuckDB SQL condition, and the chips turn into one. Once a file is open, its DuckDB database cannot read any other file.

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
  • Your Privacy Choices
  • Consumer Health Privacy
© 2026 Microsoft