Clean up a job spreadsheet before you import or share it

Find duplicate records, missing fields and changes between two job exports. Paste cells from Excel or Google Sheets, then download checked rows, a duplicate list and a readable report. No coding is needed for the spreadsheet workflow.

Best for: job-board operators, recruitment data teams and researchers who already have a job export and need to review its quality before the next import or report.

Try a small example

1. Open Input. Paste your table, including its header, into Current spreadsheet. For a first check, leave Previous spreadsheet empty and leave the advanced fields unchanged.
2. Set Maximum charge per run to $0.01 for the three-row example below, then click Start. This is a spending cap, not a fixed fee. Larger exports need a larger cap; see Pricing.
3. After Succeeded, read Result summary, Missing fields / review notes and Next step. Open Readable report for a shareable view, or Checked rows CSV to download the results.

id,source,company,title,location,url
001,example,Example Co,Engineer,London,https://example.com/jobs/1
001,example,Example Co,Engineer,London,https://example.com/jobs/1
002,example,Example Co,Designer,Remote,https://example.com/jobs/2

This synthetic example keeps two checked records and identifies one exact duplicate. Salary is missing from the input, so the tool flags it rather than inventing a value. Replace the example with your own export when you are ready.

Paste table text, not a filename or an .xlsx attachment. Common headers such as job_id, company_name, job_title and apply_url are recognized. For other flat column names, use Field mapping.

Choose the output for your task

Output	Use it to
Readable report	Review the findings without reading JSON.
Checked rows CSV	Work with normalized records and their review notes.
Duplicate groups CSV	Trace repeated records back to the input rows.
Original columns CSV	Keep all supplied current rows and columns, including duplicates and custom columns.
Original JSON	Keep the parsed input values for another workflow.

Choose a named CSV in the output selector, then click Download on its file row. To export the dataset as Excel, use Export → Excel → Download. Check the format and selected view; the export dialog may initially show JSON.

Understand the findings

• Exact duplicates: the same record identity with matching tracked fields is collapsed. The original-row export still keeps every supplied row.
• Conflicting or possible duplicates: kept for review. Two genuine vacancies can have the same title and location.
• Missing fields: identify gaps in the supplied data. The tool does not fill them with guessed values.
• Filter results: matched, excluded and unknown describe your optional filters. All retained rows remain available for review; excluded does not mean deleted.

Record identity uses source, company and ID when available, otherwise the source URL. Tracking parameters beginning with utm_ and URL fragments are removed; meaningful query parameters are retained. IDs are case-sensitive.

Compare this export with an earlier one

Paste the earlier table into Previous spreadsheet and select Both tables cover the same sources and filters only when that is true. The comparison shows which tracked fields changed. Without an earlier table, the tool performs a quality check only.

A record new to this export is not necessarily a newly posted job. A missing record is not proof a vacancy has closed. Listing missing records additionally requires both completeness settings; use them only for full, successful exports covering the same scope. The tool does not automatically fetch earlier runs.

Pricing

$1 per 1,000 checked output records ($0.001 each), plus $0.00005 per run start at supported memory settings. Platform usage is included.

The example above produces two charged output rows: $0.00205 in total. A completed run with 1,000 output records costs $1.00005 at these prices. Exact duplicates removed before output are not charged again. Flagged, excluded and unknown retained rows are charged. Previous-snapshot rows do not create a separate result charge. Reports have no separate event fee.

Set the run cap to cover the expected output count plus the start fee. Run budget too low means the tool stopped before writing result rows; raise the cap or use a smaller sample. A start fee can still apply. An interrupted run can leave charged partial output, so use a Succeeded run with OUTPUT.status: complete before importing the whole batch. A new run is charged separately.

Scope and advanced use

Up to 5,000 rows per snapshot, with a 5 MB input limit. The default memory is 512 MB and the default timeout is 60 seconds. This checks supplied exports; it does not scrape job sites, verify live vacancies or collect candidates' details.

For API workflows, use current and optional previous arrays instead of pasted tables. Clear the spreadsheet fields first. fieldMap maps canonical field names to flat source names. Optional filters include titleIncludes, locationIncludes, remoteOnly, minEmployees and maxEmployees; they are combined with AND. JSON snapshot comparisons require matching, non-empty currentScope and previousScope labels.

Your input and results remain in your Apify account's storage under its access and retention settings. Original CSV preserves parsed rows and columns, not the original file bytes. Formula-like cells are prefixed for spreadsheet safety; import numeric-looking IDs as text to preserve leading zeros.

Need help? Use the Issues tab with a small non-confidential sample, what you expected and what happened. Start with a few known records before checking a full export.