Table
Open a CSV, TSV or Parquet as a real tab on the stage and read it as a table: sort by a column, resize it, filter to the rows that mention a thing.
| Reached by | Tools → Table (CSV · TSV · Parquet) |
| Formats | CSV TSV and .tab (the extension settles the delimiter) Parquet (.parquet, .pq, .parq) |
What it does
Opens a tabular file as a real, readable table beside your sequences: sortable columns, per-column filters, resizable widths, and rows that scroll smoothly however many there are.
The data comes from your file. That sounds like it should not need saying, and it does: the point of this tool is that what you see is the file you opened, at the size it actually is, with the footer telling you how many of how many rows are on screen.
When you'd use this
Used in: Read a variant call set at a locus
- Reading the table beside the sequence it describes. A primer order sheet, a plate layout, a list of variants, a set of measurements — the thing you would otherwise open in a spreadsheet in another window and lose track of.
- In the middle of a workflow. Read a variant call set at a locus uses it: the variant table is one tab, the sequence is another, and they are on screen together.
- Checking an export. Write a CSV from another tool and open it here to confirm it says what you meant.
What this computes
Nothing is computed from your data — this is a reader, and the honesty of a reader is in what it tells you about itself:
- the rows on screen and the rows in the file, stated in the footer as a ratio. A preview reads the head of a file; a table showing the first two hundred of four million rows must never present itself as the file
- the column count and row count of the sheet being shown
- the sheets a document contains. A CSV, TSV or Parquet file is one sheet; a workbook is not, and the sheet list is always produced so a multi-sheet document cannot hide its other sheets behind no affordance at all. The picker appears exactly when there is more than one thing to pick
- the delimiter, settled by the extension rather than sniffed —
.tsvand.tabare tab separated,.csvis comma separated
Where it lives
Tools → Table, and by opening a tabular file from the file manager or the sidebar — which is the usual route. It draws as a stage tab beside your other documents, and in the file manager's preview aside.
Inputs and outputs
Accepts CSV, TSV and .tab, Parquet in its several extensions, and spreadsheet workbooks.
Produces no file. This is a view; nothing is written back, and the table is read-only.
A file it cannot parse opens in the text viewer instead, where you can see the bytes — which is more useful than a table-shaped error, because the usual cause is a file that is not the format its extension claims.
Demo
There is no screencast. Open any .csv in a project and it draws immediately.
Worked example
Follow the Read a variant call set at a locus workflow. Its variant table opens here beside the sequence, and the first thing worth doing is sorting by position — which is how you find out whether the file is sorted, which determines whether an index would help.
Then filter to the rows mentioning a gene name. The footer's row ratio updates, so you can see both how many rows matched and how large the file you are filtering actually is.
Controls
| Control | What it does |
|---|---|
| Column sort | click a header; click again to reverse |
| Per-column filter | narrows to rows matching in that column |
| Column resize | drag a header edge; useful when one column carries long identifiers |
| Sheet picker | appears when the document has more than one sheet |
In the file manager's narrow preview the controls and the row and column counts are drawn by the host around the table rather than stacked inside it, so the table keeps the width.
Algorithms you can choose
None — parsing a delimited file is a specification, not a choice. The one decision the reader makes is the delimiter, and it is taken from the extension rather than guessed, because sniffing gets a file with commas inside quoted tab-separated fields wrong in a way that looks plausible.
The science
There is no science here; this is a file reader. The one convention worth stating is that the first row is treated as a header. A file whose first row is data will show that row as column names — the sort of thing that is obvious once seen and invisible if you are skimming.
Numeric-looking columns are shown as they appear in the file rather than reformatted, so a leading zero, a thousands separator or a locale-specific decimal comma survives to be looked at rather than being silently reinterpreted.
Limitations
- A preview reads a head of the file, not all of it. The footer says so with the real ratio. Open the file as a full tab for the whole thing.
- Read-only. There is no editing, no formula evaluation, and no writing back.
- No type inference and no aggregation. No sums, no group-by, no charts. A column of numbers is a column of text as far as this reader is concerned — sorting is lexical unless the column is genuinely numeric.
- Very wide tables scroll horizontally rather than reflowing; there is no column hiding.
- A workbook's formatting is not shown — values only, no cell styling, no merged cells.
Compared with other tools
The comparison that matters is with a spreadsheet application, not with another plasmid editor: a spreadsheet does far more with a table than this does, and does it in a different window. What this offers instead is the table in the same workspace as the sequence it describes, in a tab beside it, with an honest row count. When you need to compute over the data rather than read it, the spreadsheet — or the built-in Python and R consoles — is the right tool, and this one does not pretend otherwise.
Export
Nothing is exported from this view. The file is already a file; it is in your project and can be downloaded from the file manager unchanged.
API
There is no engine behind this surface. It is the repository's shared data-table drawing a parsed file, so the programmatic equivalent is parsing the file yourself — which is what the Python and R consoles are for, and where a table you want to compute over should go.
Automation
There is no workflow node for viewing a table. Pipelines that produce tabular results write real files into the project, and those files open here — which is the hand-off: the node writes, this reads.
Provenance
CSV and TSV are read per their usual conventions, with the delimiter taken from the extension. Parquet and workbook files are read by their format specifications. Round-tripping is not claimed here because nothing is written.
Related
The file manager, where tabular files are listed and previewed. The variant table, which is the right view for a VCF specifically. The Python and R consoles, for computing over a table rather than reading it.
Screenshots
No frame is committed for this surface yet. What one would show: a real CSV or Parquet table with its filter, its column types and its row count.
The absence is stated rather than left as a blank heading: a picture that does not exist yet is worth knowing about, and a placeholder would be worse than either.