Home / GIS and Open Map Data for Beginners

The QGIS Attribute Table: Rows Are Places, Columns Are Facts

Overhead view of a team analyzing graphs and charts on laptops

Photo: olia danilevich / Pexels

Click any feature on a QGIS map and the program can tell you things about it, because behind every shape sits a row. The attribute table is the layer's other face: one row per feature, one column per fact, and the whole analytical life of the dataset flowing between the two views. Sort it and the extremes of your data announce themselves. Filter it and a question becomes a subset. Select in it and the map lights up with the answer. Most beginners treat the table as a place to peek at a value; professionals treat it as the primary instrument, the map's control room. This guide teaches the table as an instrument: how to open it in its three useful scopes, how its two views serve different work, how sorting and filtering isolate the records a question is actually about, how selection in the table and selection on the map are the same act in two windows, and how editing and the field calculator extend it from reading to writing. It pairs with our field calculator guide, which takes the calculating further, and with select by location, which asks the spatial questions the table cannot.

Interactive tool

Rows Are Places: The Attribute Table Checklist

Check off each step as you finish it. Your progress stays on this page.

0 of 7 complete

0% done

Open It Three Ways: All, Selected, Visible

The table opens from the layer: right-click the layer in the panel and choose Open Attribute Table, or select the layer and press F6. That gives you every row in the dataset. Two modifier variants earn their place in your habits immediately. Shift+F6 opens the table showing only the selected features, which is the view you want after any selection operation, when the question is what exactly did I just catch. Ctrl+F6 opens it showing only the features visible in the current map view, which turns the table into a local instrument: zoom to a neighborhood, open the visible-only table, and you are reading the data of that neighborhood alone. The header of the table reports the counts as you go, total features, filtered, selected, and those three numbers, watched consciously, prevent the classic table accident of acting on a subset while believing you acted on the whole. Learn to glance at them the way a pilot glances at instruments: before every operation, ask which rows am I really looking at right now.

Once open, the table can also change its own scope without reopening: buttons along its top switch the view between all features, the selected ones, and the visible ones, and a further control decides whether selecting a row also zooms or pans the map to it. Set that behavior deliberately for the job in hand. Auditing a dataset wants the row and the map side by side with no jumping; hunting one record's location wants the map to fly to every row you touch. The table is not one tool but a small console of them, and the console's settings, chosen consciously at the start of a session, are the difference between an instrument that answers and one that merely displays.

The Two Views: Table for Many, Form for One

The table has two faces, switched by the icons in its corner. The table view is the spreadsheet face: all rows, all columns, built for comparison, sorting, and scanning, the view in which patterns and outliers are visible. The form view shows one feature at a time, its fields arranged as a form with navigation arrows to step through records, which is the right face for reading or editing a single feature's full record, especially in layers with many columns where the table view scrolls forever sideways. The form view also respects the layer's configured field widgets and layouts, so in a well-built project it presents the record the way its designer intended, with dropdowns and grouped sections rather than raw columns. Neither view changes the data; they are two chairs in front of the same desk. Skilled work alternates: find the interesting records in the table view, where the crowd makes the anomaly visible, then read or edit the chosen one in the form view, where its whole record fits on one screen. Knowing which chair a task wants is a small thing that saves a large amount of scrolling.

The form view has one more talent the table cannot match: it is the face a well-designed project shows its field crews, with fields ordered, grouped, and widgeted for the work, dropdowns where values must be chosen, calendars where dates belong. When you meet a layer whose form view feels designed, someone cared about the people entering the data, and the data's quality usually shows it. When you build the layer, that care is yours to give, and the attribute table's form settings are where you give it.

Man writing in a notebook while using a laptop and calculator at a desk
Man writing in a notebook while using a laptop and calculator at a desk

Sort and Filter: Isolate the Question

Sorting is the table's fastest analysis. Click a column header to sort by it, click again to reverse, and the extremes of any numeric field, the largest parcels, the oldest records, the highest readings, rise to the top or bottom where you can meet them. For anything fancier, the sort dialog accepts an expression, so you can sort by a computed value, a ratio of two columns, a name's length, without altering the data. Filtering answers the sharper question: show me only the rows that match. The form filter, opened with Ctrl+F, lets you build conditions field by field, this district, that status, values above a threshold, with the matching rows alone surviving in the table and the filter count in the header proving how many. The advanced filter takes an expression instead, for conditions that span fields and functions. Two disciplines keep filtering honest. First, a filter is a lens, not an edit: filtered-out rows still exist, still export if you are careless, and still count in operations that ignore the filter, so know which of your tools respects it. Second, filters compose with selection and with the map in ways that reward a moment's thought before a click. The table rewards the user who treats every view of it as a claim about which rows are in play.

Select From the Table: The Map Answers Back

Selection is where the table and the map reveal themselves as one instrument. Select rows in the table, by clicking, shift-clicking a range, or running select by expression for anything rule-based, and the corresponding features highlight on the map instantly; select features on the map and their rows highlight in the table. This two-way mirror is the core analytical gesture of GIS: ask in whichever medium the question is phrased, read the answer in the other. The expression route is the powerful one. Select by expression lets a sentence become a selection: all parcels in this district larger than that size with this zoning, all readings above the threshold taken this year. The selected set can then be zoomed to, counted, exported, or promoted into a layer of its own, which is the subject of our companion guide to making a new layer from a selection. The habit to build is narrating the selection before making it: say which rows the question wants, in words, then write the expression that matches the words, then check the count the table reports against your expectation. When the count surprises you, the expression or the data is teaching you something, and either lesson is worth more than the selection.

Edit and Calculate: Writing in the Table

The table is read-only until you say otherwise, and the saying is Toggle Editing, Ctrl+E or the pencil, after which cells accept changes directly. Editing in the table is the fastest way to fix the small stuff: the misspelled name, the misclassified row, the missing value you just learned. It is also, being fast, the fastest way to damage data at scale, so the disciplines are simple and strict: edit deliberately, save consciously, and remember that closing without saving discards, which is a feature when you have made a mess. The field calculator, Ctrl+I, is the table's power tool: it writes a whole column at once from an expression, updating an existing field or creating a new one, computing areas, ratios, categories, and cleaned versions of messy text across ten thousand rows in the time a manual edit takes for ten. Our field calculator guide is devoted to it. One more quiet feature deserves discovery: conditional formatting, which colors cells by rule, values above a limit in one color, missing values in another, turning the table into a data-quality dashboard where the problems glow. For any dataset you inherit, an hour in the table with sorting, filtering, and formatting will teach you more truth about it than a day of staring at its map, because the map shows you the data's geography and the table shows you its character.

The Audit Hour: Counts, Nulls, and Impossible Values

Every inherited layer deserves one hour in its table before it is mapped, analyzed, or believed, and the table's tools make that hour systematic. Start with the counts: the feature count against what the source claims, the count of each category in the classifying field, because a dataset that says it holds a state's wells and holds a county's has told you its first truth. Hunt the nulls next: filter each important field for emptiness, since missing values cluster where collection failed, and a map drawn over unexamined nulls is a portrait of the gaps wearing the data's clothes. Then the impossible values, found by sorting: the negative areas, the dates from the eighteenth century, the percentages in the thousands, each one a small confession about how the data was made. Finish with duplicates, the same name or ID appearing twice where the world holds one, visible the moment the table is sorted by the field that should be unique. None of this requires a single spatial operation; it is reading, with the table's instruments, and it routinely changes what the project does next, which questions the data can answer and which it must decline. The map will get its turn. Data that has survived its audit hour makes a better one, because every layer is two things, and the table is the half that tells you whether the other half deserves your trust.

Related Guides

Readers working through this topic usually need its closest companions too. These GIS and open map data guides cover the neighboring ground:

Frequently Asked Questions

What is the difference between F6, Shift+F6, and Ctrl+F6?

F6 opens the attribute table with all features. Shift+F6 opens it showing only the currently selected features. Ctrl+F6 opens it showing only the features visible in the current map extent. The header counts always tell you which scope you are in and how many rows each category holds.

Does filtering the table delete the other rows?

No. A filter only changes which rows the table displays; every feature remains in the layer. Be careful, though, because some operations act on all features regardless of the filter, while exports and selections can be scoped. Always know whether your next action respects the filter before you take it.

How do I select features with a rule instead of clicking?

Use select by expression in the table toolbar or the main selection tools: write an expression that describes the rows you want, such as a district combined with a size threshold, and the matching features select in both the table and the map. Check the reported count against your expectation before proceeding.

Can I calculate a new column from existing ones?

Yes, with the field calculator, Ctrl+I. It writes an entire field from an expression, creating a new field or updating an existing one, across all rows or only the selected ones. It is the standard way to derive areas, ratios, categories, and cleaned text at scale.

Why can I not type into the table cells?

The layer is not in edit mode. Toggle editing on with Ctrl+E or the pencil button, make your changes, then save the edits and toggle editing back off. The read-only default is a safety feature that protects layers from accidental changes.

Photo credits: olia danilevich (Pexels); Mohamed hamdi (Pexels). Photos are served from this site; follow the credit to the library page for the original.