Using the GUI

← Contents

• The window
• Opening files
• Moving around and selecting
• Copying
• Going to a row, cell or column
• The sidebar
• Sorting
• Filtering
• Searching
• Column statistics
• Histograms
• Scatter plots and 2D density
• Group By
• Hiding and freezing columns
• Column types and formats
• Sheets
• Running SQL
• Exporting and converting
• Editing cells
• Reloading changed files
• Display options
• Menu reference

On macOS, use Cmd wherever this manual says Ctrl.

The window

• Menu bar: File, Edit, View, Window and Help. On macOS the menus are in the system menu bar.
• Column headers: the column name, a sort button (the small arrows on the right) and, when the column is filtered, a filter mark. A small mark on the left tells what the column seems to hold: # numbers, @ dates, ? booleans (see Column types and formats).
• Row numbers: the position of the row in the current view (after sorting and filtering).
• Sheet tabs at the bottom: one tab per sheet. A filtered sheet shows (filtered) after its name.
• Sidebar on the right (hidden at first): the Content, Search, Filter and Statistics panes.

Each file opens in its own window. The Window menu lists the open windows so that you can switch between them.

Opening files

• File → Open... (Ctrl+O), or the Open... button of an empty window. The file dialog lists every supported format.
• File → Open Recent lists the last ten files opened, newest first, in every window. Clear Recent Files empties the list. A file that cannot be opened any more is taken off it.
• Give files on the command line: tableskim a.csv b.xlsx.
• macOS: drop files on the application icon, or use Open With in the Finder.

A file opened from a window that already shows one opens in a new window. File → Close (Ctrl+W) closes the window, File → Quit (Ctrl+Q) quits TableSkim.

Rows are read in the background: the table appears at once, cells that are not loaded yet are filled in as they arrive, and the window stays responsive.

Moving around and selecting

• Scroll with the mouse wheel or the scroll bars, or move the current cell with the arrow keys, Page Up / Page Down, Home / End.
• Click a cell to select it; drag or Shift+click to select a range.
• Click a row number to select the row, or a column name to select the column.
• Drag the border between two column headers to change the width of a column.

Copying

Edit → Copy (Ctrl+C) copies the selected cells as tab-separated text (TSV), which can be pasted into a spreadsheet. Hidden columns are left out. Tabs and line breaks inside a cell are replaced with spaces, and a cell that starts with a double quote is quoted so that it pastes back as it was. If some selected rows are not loaded yet, they are skipped and a message says how many rows were copied.

Going to a row, cell or column

Edit → Go To... (Ctrl+G) jumps to:

• a row number, such as 120;
• a cell, as B12 or R12C2;
• a column, by its letter (AB) or by its name (Salary). A column name is matched exactly first, then ignoring upper and lower case, and wins over a column letter.

The sidebar

View → Show sidebar shows or hides the sidebar. It also appears automatically when you use Find, Filter or Column Statistics. Drag its left edge to make it wider.

The Content pane shows the whole value of the selected cell, wrapped to the width of the pane. This is useful for long text. Double-click a cell to open it.

Sorting

Click the sort button (the small arrows at the right of a column header) to sort by that column: the first click sorts in ascending order, the second in descending order, and the third restores the file order. The arrow in the header shows the current direction. Sorting by one column replaces any other sort.

Right-click a column header for the column menu, which has Sort Ascending, Sort Descending and No Sort as well as filtering, statistics, histograms, hiding and freezing.

How values are compared (numbers, dates, text, empty cells) is explained in Sorting, Filtering, Searching and Statistics.

Filtering

Chart selections are kept as Additional narrowing, ANDed with the conditions in the Filter pane even when Match: Any is selected. Editing those conditions keeps the chart selection. Click Additional narrowing (...) - Clear to remove only the chart selection; Clear removes both. Further chart selections narrow the remaining rows again.

Quick filters

Right-click a cell and choose Filter: = "value" to show only rows with the same value in that column, or Filter: ≠ "value" to hide them. Each quick filter is added to the current filter as another condition.

The Filter pane

Edit → Filter... (Ctrl+Shift+F), or Filter by This Column... in a column or cell menu, opens the Filter pane.

• Match: All / Any: show rows that match all conditions (AND) or any condition (OR).
• Each condition has a column, an operator and a value. The × button removes the condition.
• + Add condition adds another condition.
• Apply (or Enter in a value field) applies the filter. The line below shows the result, for example "4 rows of 54".
• Clear removes all conditions.

The operators are contains, not contains, =, ≠, <, ≤, >, ≥, is empty, is not empty, matches and not matches (regular expressions), is a number and is a date. Comparisons follow the type of each cell: numbers compare as numbers, dates as dates, and = 2024-01-31 matches any time on that day. See Sorting, Filtering, Searching and Statistics for the details.

Edit → Clear Filter, or Clear Filter in a column or cell menu, removes the filter.

Searching

Edit → Find... (Ctrl+F) opens the Search pane. Type the text and press Enter or Search.

• Matching cells are highlighted in the table, and the list shows the row, column and value of every hit. Click a hit to jump to it.
• The search ignores upper and lower case and finds the text anywhere in a cell.
• Check Regex to search with a regular expression (RE2 syntax).
• The search runs in the background; the status line shows the progress and the number of hits.

Column statistics

Edit → Column Statistics (Ctrl+Shift+I), or Column Statistics in a column or cell menu, opens the Statistics pane for the current column. Choose another column from the list at the top.

• The summary shows the number of rows, empty cells, distinct values and, for numeric columns, numbers, min, max, sum, mean and standard deviation.
• The table below lists the most frequent values with their counts and percentages.
• Statistics cover the rows of the current view, so they follow the filter.
• Double-click a value to filter by it. Right-click a value for Filter: =, Filter: ≠ and Copy Value.
• Drag the border between the summary and the list to resize them.
• Cancel stops a computation that takes long on a large file.

Histograms

Charts and Group By shows, with examples, which chart to use and how to read it.

Edit → Histogram... (Ctrl+Shift+H), or Histogram in a column or cell menu, opens a histogram of the current column in a window of its own. Choose another column from the Column list at the top; the line next to it summarizes what was counted, for example "53 numbers in 30 bins, 1 empty".

• Numbers are counted into about 30 bins of round widths, such as 50000 or 0.2. Integer columns whose values span at most about 30 numbers (such as ages 20 to 45) get one bar per value.
• Dates and date-times are counted into bins of whole days, weeks, months, years and so on, depending on the range of the column.
• Other columns (text) show their most frequent values as horizontal bars instead.
• Point at a bar to see its range and count below the chart, for example [600000, 650000): 4 (7.5%). A range includes its lower bound and excludes its upper bound ([ and )), except the last bar of numbers, which includes both ([1450000, 1500000]).
• Double-click a bar to filter the sheet to the rows of that bar. The histogram is then recomputed for the remaining rows, so you can drill down step by step. Edit → Clear Filter goes back to all rows.

The histogram always follows the current view of the sheet: changing the filter, in the Filter pane or anywhere else, recomputes it, and the summary says "(filtered rows)". Below, the 1994–1995 bar was double-clicked: the sheet now shows those six rows and the histogram shows them in monthly bins.

For a text column, each bar is one value; double-clicking it filters the sheet to that value.

How bins are chosen is described in Sorting, Filtering, Searching and Statistics.

Save Image... writes the chart as a PNG or SVG file (by the extension of the name, or the type chosen in the dialog). The image is the chart as shown, at its size in the window, with the column name under the axis.

Scatter plots and 2D density

Edit → Scatter Plot... (Ctrl+Shift+P), or Scatter Plot in a column or cell menu, plots two columns against each other in a window of its own. If two or more columns are selected, the first and last are X and Y. Otherwise TableSkim picks columns holding numbers or dates (judged by the rows around the current cell): Y is the current column, or the nearest such column, and X the nearest such column to its left (or, if there is none, to its right). Choose other columns from the X and Y lists, and the kind of plot from Show:

• Points draws a point per row. Many points on top of each other look darker, so dense areas stand out. A large table is plotted as a random sample of 200,000 rows (the summary says so); the sample is the same each time.
• Density divides the plot into a grid of about 50 × 50 cells and colours each cell by the number of rows in it, from dark purple (few) to yellow (many), with the scale on the right. Log color scale (on by default) keeps cells with a few rows visible next to very dense ones. Empty cells stay blank. The density always counts every row.

Y must hold numbers or dates. X may hold numbers or dates too, or text (see Text on X: box plots below). Rows where either value is empty or of another kind are left out and counted as "skipped" in the summary.

The line under the summary shows how closely the columns are related, for example r = 0.0945, R² = 0.0089, y = 2641.23 x + 744351: Pearson's correlation coefficient r (from −1 to 1), its square R², and the least-squares line. They are computed over every row plotted, not only the sample of points. Regression line draws that line on the plot. The equation is left out when a column holds dates, whose values would be in microseconds.

• Point at a point to see its values and row number below the plot; click it to move the current cell of the sheet to that row.
• Point at a cell of the density to see its ranges and count, for example x [-30000, -25000) × y [60, 65): 22 rows (0.2%); double-click it to filter the sheet to the rows of that cell.
• Drag a rectangle (in either view) to filter the sheet to the rows whose X and Y values lie inside it.

Like the histogram, the plot follows the current view of the sheet, so after filtering it is drawn again for the rows left. Edit → Clear Filter goes back to all rows. See Sorting, Filtering, Searching and Statistics for the details.

Text on X: box plots

When the X column holds text (such as a city or a department), each of its values is a category with a column of its own: the 30 most frequent values, most frequent first, and (other) for the rest. Show then switches to Box Plot (it is available only then):

• Each box spans the first to the third quartile of the category's Y values, with the median as a thick line and the mean as a red cross. The whiskers reach the furthest values within 1.5 times the box's height from it; values beyond them are drawn as small rings.
• Point at a box to see the category's count, minimum, quartiles, maximum, mean and number of outliers below the plot; double-click it to filter the sheet to that category.
• Density shows a histogram of Y for each category, side by side, as coloured cells: double-click a cell to filter to that category and range.
• Points spreads each category's points sideways over its column, so that they do not hide each other.

There is no correlation or regression line for categories, and no rectangle to drag.

Saving the plot

Save Image... writes the plot as a PNG or SVG file, as shown: the kind of plot, the colour scale, the regression line and the axis titles (the column names). The type follows the extension of the name (.png or .svg), or the type chosen in the file dialog. SVG files keep lines and text sharp at any size; the points of a scatter plot are an image inside them.

Group By

See also Charts and Group By for worked examples.

Edit → Group By... (Ctrl+Shift+G), or Group By… in a column menu, counts the rows of the current sheet by the values of some columns, like a pivot table or SQL's GROUP BY, and shows the result in a new sheet named Group By.

• Group by: rows with equal values in these columns form a group. The selected columns (or the clicked one) are checked at first. With none, every row is in one group, which gives totals.
• Aggregate: columns to summarize in each group, with the functions checked below: count (non-empty values), distinct (distinct values), sum, mean, min, max and std dev (sample standard deviation).

The result has one row per group, in sort order of the group's values: the group columns, count (the rows of the group), and a column such as Salary sum for each aggregated column and function. It covers the rows of the current view, so it follows the filter.

The Group By sheet can be sorted, filtered, exported and grouped again like any other. Grouping again replaces it. It is a snapshot: it does not change when the original sheet does, but stays when the file is reloaded. Empty cells form a group of their own; how values are compared is described in Sorting, Filtering, Searching and Statistics.

Hiding and freezing columns

Hiding columns

Right-click a column header and choose Hide Column. If several columns are selected and you right-click one of them, Hide Columns hides them all. View → Show All Columns (or Show All Columns in the column menu) shows them again.

Hidden columns are skipped when moving the cursor, copying and exporting. They still take part in sorting, filtering and statistics.

Freezing columns

Right-click a column header and choose Freeze Columns up to Here: that column and all columns to its left stay in place when you scroll sideways. Unfreeze Columns in the column menu, or View → Freeze Columns, turns it off. At least one column must remain unfrozen.

Column types and formats

TableSkim looks at the first 1,000 rows read to tell what each column holds, and marks the header: # for numbers, @ for dates and date-times, ? for booleans (TRUE / FALSE). Text and mixed columns have no mark. Text cells count as what they read as, so a CSV column of 42 and 3.5 is a number column. A few stray values (a - among numbers) do not change the kind, as long as 90% of the values agree.

Format in the column header menu changes how the column shows its values; the first line of the submenu shows its type, for example "Type: Number (whole numbers)":

• Thousands Separator: 1234567.5 is shown as 1,234,567.5.
• Decimal Places: As Stored, or 0 to 6 places, rounded (1234.5678 with 2 places is 1234.57).
• Date Format: 2024-01-31, 2024/01/31, 01/31/2024, 31/01/2024, 2024年1月31日 or Jan 31, 2024. A time of day stays as it is.
• Reset Format goes back to the values as stored.

If several columns are selected and you right-click one of them, the change applies to all of them. Numbers and dates held as text are formatted too. Only the display changes: sorting, filtering, copying and exporting use the values as they are. Formats belong to the sheet in this window and are kept when the file is reloaded.

Sheets

Excel workbooks and databases show one tab per worksheet, table or view at the bottom of the window. Click a tab to switch. Each sheet keeps its own sort, filter, hidden and frozen columns.

Running SQL

For DuckDB, SQLite, Parquet and JSON files, Edit → Run SQL... (Ctrl+E) opens the SQL window. Type a query and press Run (Ctrl+Enter); the result appears as a new sheet named SQL Result.

The SQL window has a history (History, Ctrl+Up / Ctrl+Down) and completion (Tab). See Running SQL for details.

Exporting and converting

• File → Export View... (Ctrl+Shift+E) saves the current sheet as shown (sorted, filtered, without hidden columns) to CSV, TSV, Excel or Parquet.
• File → Convert to Parquet / DuckDB... converts a text file (CSV, TSV, VCF) to Parquet or a DuckDB database, for fast access later.

Both run behind a progress dialog with a Cancel button. See Exporting and Converting.

Editing cells

Cells of CSV / TSV and Excel files can be changed in edit mode only: Edit → Edit Mode (Ctrl+Shift+M) turns it on and off, and the title shows [Editing]. Then F2, Enter, typing or a double click edits a cell, Delete clears the selected cells, Edit → Paste (Ctrl+V) pastes tab-separated text, Edit → Undo / Redo take changes back, and File → Save (Ctrl+S) or File → Save As... (Ctrl+Shift+S) writes them. Edited cells are green until saved. See Editing Cells.

Reloading changed files

• File → Revert reloads the file from disk.
• File → Auto Revert (on by default) checks the file regularly (every second while the window is active, every five seconds otherwise) and reloads it when it has changed. This is handy for watching a file that another program writes.
• File → Follow End of File (Ctrl+Shift+T) keeps showing the end of a file that grows, like tail -f: the file is checked twice a second (whether or not Auto Revert is on), and the current cell moves to the last row whenever rows are added. For CSV / TSV files that are not compressed (in UTF-8 or Shift_JIS), only the appended lines are read, so even a large log file keeps up; a line being written is shown once it ends. Other files, and files that were rewritten rather than appended to, are reloaded. A sorted or filtered sheet sorts and filters the new rows too. Choose the item again to stop following.
• File → Show File in File Browser opens the folder of the file in the system file manager.

After a reload, a SQL result sheet is computed again by running the last query. With unsaved edits, Revert asks before dropping them, and Auto Revert waits (the title shows (changed on disk) if the file changes).

Display options

The View menu:

Items that do not apply to the current file (for example the header option for a Parquet file) are disabled.

Menu reference

Help → User Manual (F1) opens this manual in a window of its own, in Japanese if that is the language of the system and in English otherwise; the list at the top right switches the language. Back / Forward (Alt+Left / Alt+Right) return to pages seen before, and Contents goes to the table of contents. The manual is built into TableSkim, so it works offline.

Help → About... shows the version and the licenses of the bundled third-party software and fonts.