How do I keep data in an online table instead of passing Excel and CSV files around?
Upload a CSV or Excel of up to 2 GB to Sealique: the rows become a database table shared with roles, edited in the browser and read over an API.
Spreadsheets travel by e-mail, get renamed v3-final, and nobody knows which copy is right. A Sealique data table keeps one copy in a database: you share it with viewers, contributors, editors and managers, collect new rows through forms, clean the data with rules and scripts, analyse it with charts, pivots and dashboards, and read it from other systems over an API.
Create a table Create an account Verify a signed document
Key facts
- A data table accepts CSV and Excel uploads of up to 2 GB, imported in the background in chunks.
- Rows are stored in a database, not as a file, so several people edit the same table at the same time without copies; charts, pivots and dashboards read the same live rows.
- Every change is recorded with who did it and when; named stamps let a manager roll the table back to a checkpoint.
- Forms built from the table's fields collect rows from assigned people, a link or the public, with dropdowns that depend on a parent choice.
- A read and write API with per-table keys exposes the rows as JSON, with filters, search and paging.
Who are data tables for?
Data collectors who gather records in the field and need one place where entries land instead of merging dozens of spreadsheets. Each collector gets a form or a contributor role and sees only what they added.
Analysts who receive files from many sources and spend their time reconciling versions. One table with a change history replaces the chain of e-mailed copies, and charts and pivots run on the live rows.
Teams that keep a shared register, such as members, assets, suppliers or beneficiaries, and need roles: some people only read, some add, some edit, a manager controls the structure.
How do imports work?
Files are uploaded in chunks, so a 2 GB CSV or Excel workbook does not depend on one fragile request. The import runs in the background: you see the number of rows done and any lines that were skipped, with the reason.
Columns are mapped to fields by name; you can change the mapping before the import starts. Types are detected from the values (whole numbers, decimals, dates, yes/no, repeated values as choices) and you can correct them.
A preview then checks the first 500 rows against the mapping and the table's rules exactly as the import will, without writing anything: rows with problems are listed per column with the reason, so you can fix the file or the rules first.
Imports are recorded as one change batch, so a wrong file can be undone in one step.
What do cleaning rules do?
Rules check every new or edited row: required, unique, pattern, length, numeric sign and bounds such as greater than 5 or at most 2 decimals, date range, one of a list of options, exists in another table, no duplicates on a combination of fields. A rule is an error that blocks the row or a warning that is shown and kept.
Clean-up actions fix existing rows in bulk: trim spaces, change case, normalise numbers and dates, fill empty cells, find and replace. Each action is one batch in the history and can be rolled back.
Quality scripts are small read-only Python scripts that read the rows and flag the ones that look wrong, for checks a rule cannot express. They cannot change data.
How do forms collect rows?
A form is built from the table's fields: pick which ones to show, give them a label and help text, choose a widget (text, number, date, dropdown, radio buttons, checkboxes, toggle) and mark the required ones.
Dropdown options are typed in or taken from the values already stored in the column, and a column whose values repeat is suggested as a dropdown. Dropdowns can depend on another field: choose a province and the district list narrows. A column linked to another table offers that table's values. Entries are checked against the table's rules before they are saved.
Access is per form: assigned people sign in and find the form under Forms, anyone with the link needs no account, public forms are open to everyone. A QR code of the link is ready to print.
What does the API offer?
Each table can have several API keys with a read or write scope. The read API lists rows as JSON with filters per field, full-text search, sorting and paging, and returns the field definitions.
The write API adds, updates and deletes rows through the same rules as the interface, and every change is attributed to the key in the history.
How does versioning work?
Every insert, edit, delete, import, clean-up and schema change is an entry in the table's history with the actor and the time. A row's own history shows each version of its values.
A stamp names a point in the history, for example before an import. A manager can roll back to a stamp or undo one batch; the rollback is itself a recorded change, so nothing is lost.
Data tables, step by step
Create a table
Sign in, open My tables and choose New table. Upload a CSV or Excel file, paste rows, or start from a blank table and define the fields yourself.
Check the fields
The import proposes a type for each column (number, date, choice…). Fix the ones it got wrong, mark required or unique fields and add dependent choices such as province and district.
Preview before importing
The first 500 rows are run through the mapping and every rule of the table before anything is written: you see which rows pass, which columns cause problems and why, then start the import in the background.
Share it with roles
Invite people by e-mail as viewer, contributor, editor or manager. Someone without an account is attached automatically when they sign up with that address.
Collect rows with a form
Build a form from the table's fields, turn columns with repeated values into dropdowns (the options come from the values already in the table), set rules such as greater than 5, decide who can fill it in (assigned people, anyone with the link or the public) and share the link or its QR code. Every entry becomes a row.
Clean and analyse
Add rules (required, unique, pattern, range, one of, exists in another table), run quality scripts whose results sit on the Data quality page, and build charts and pivots from the Dashboards area. Put the best of them on a dashboard with KPI cards and a written summary, then publish it on a public link or export it to PDF or PowerPoint.
Connect other systems
Create an API key with read or write scope and fetch the rows as JSON with filters, search and paging, or push new rows from another application.
Data tables is included with a Sealique account; the processing happens in your browser and on our servers, nothing is shared with third parties.
What Data tables includes
- CSV and Excel import up to 2 GB, with column mapping, type detection and a checked preview of the first 500 rows
- Typed fields: text, whole and decimal numbers with sign and bounds, dates, choices, dependent choices, e-mail, phone, links to other tables
- Relationships between tables: a column must match a column of another table, forms offer its values and linked rows cannot be deleted by mistake
- Roles per table: viewer, contributor, editor, manager; contributors see only their own rows
- Change history with stamps and rollback
- Forms with dependent dropdowns, assigned people, link or public access and QR codes
- Cleaning rules and read-only quality scripts that flag problems
- Charts, pivots and column statistics computed live
- Dashboards with KPI cards, charts, pivots and text blocks on a drag-and-drop grid, with tabs, filters, a public link and PDF or PowerPoint export
- Read and write API with per-table keys
- CSV and Excel export
Questions people ask
Can I replace Excel with an online table for my team?
Yes. Upload the workbook once; the rows live in a database table that everyone opens with their own role, so there is one copy instead of many files.
Export to CSV or Excel at any time if a file is still needed for a report or another tool.
How large a file can I upload?
CSV and Excel files of up to 2 GB; the upload is chunked and the import runs in the background with progress and skipped-line reports.
Very wide sheets work too: every column becomes a field, and you can skip columns in the mapping.
Can people add rows without an account?
Yes. A form with link or public access needs no account; assigned forms ask people to sign in so their rows are attributed to them.
Rules still apply to form entries, so a wrong district for the chosen province is refused with a clear message.
What happens if someone imports the wrong file?
An import is one batch in the change history. A manager undoes it in one step or rolls the table back to a stamp taken before it.
Stamps are quick to create; take one before any big change.
Can another application read the data?
Yes. Create an API key with the read scope and fetch rows as JSON with filters, search and paging; a write scope allows adding and editing rows.
Keys are per table and can be revoked at any moment.
Can I link two tables?
Yes. Relationships in the table settings link a column to another table's column: values must exist there, forms offer them, and linked rows cannot be deleted while in use.
Removing the relationship turns the column back into plain text and keeps the stored values.
Can I build a dashboard from a table?
Yes. The chart and pivot builder draws them; dashboards combine them with KPI cards and text blocks, publish on a public link and export to PDF or PowerPoint.
Interpret writes a plain-language summary of a widget's numbers, and related tables can be combined through their links.
Do quality scripts change my data?
No. Scripts run in a read-only sandbox: they can read rows and flag them, but cannot write, import modules, open files or reach the network.
Each run keeps its flags with a link to the row, and a time limit stops runaway scripts.