Moving a spreadsheet in five steps

Set aside an afternoon and work on a copy of the spreadsheet, never the original.

  1. Check the numbers. Sort by object number and look for blanks, duplicates and numbers typed with spaces or the wrong separator. Fix them in the copy.
  2. One row per object. Make sure each object has its own row. If accession details sit on a separate sheet, copy the donor and date columns onto the object rows or keep an accession number column.
  3. Save as CSV. Save the copy as CSV, comma separated, with the column names in the first row.
  4. Match and check. Import the file, check each column's match, choose how dates are written, and read the report of skipped rows and cells to check.
  5. Import and save the file. Import, look through a few records, then save the catalogue file so the imported catalogue has a copy outside the browser.

Numbers decide where every row goes

Each row is filed under an accession. If the sheet has an accession number column, that is used; if not, the accession is the object number without its last part, so 2019.4.2 goes under 2019.4. A row whose number is not year based (Box 7, or blank) cannot be placed and is reported with its spreadsheet row number, for you to number and import again.

Numbers are matched by value, so 2019.4.2 and 2019.004.002 are the same number. A row counts as already in the catalogue only when an object with its number also has its object name and description, as when the same file is imported twice. A row that shares only the number is a different object, such as a second thing that was given a number already in use, and it is imported; the report gives its row number and the Numbers tab lists the pair as a duplicate to renumber with a note. In the example below, row 6 repeats a row exactly and is left alone, and row 7 (2019.4.2) is imported as a second object with that number.

Cells the import reads for you

  • Dates written 6/2/2019, 2 June 2019, 2019-06-02 or as an Excel day number; the order of day and month is a choice you make, because 03/04 is a different day on each side of the Atlantic.
  • Dimensions in one cell, such as 40 x 25 x 12 cm, split into height, width, depth and unit.
  • A location in one cell, such as Parlor / Mantel, split into room, case and shelf.
  • Acquisition words such as donation, bought or bequeathed, read as a method.
  • Condition words such as good or poor; anything else is kept in the condition notes and listed to check.

Export works the other way: CSV for a spreadsheet or for a move to a larger system later, with every cell that a spreadsheet would run as a formula written as text.

Worked example: a five row sheet

This sheet has two gifts, a bequest entered twice, and a row of loose photographs with no proper number.

How the import matched each column
Column in the sheetMatched to
Catalog NumberObject number
Object NameObject name
DescriptionDescription
DonorSource or donor
Date ReceivedDate received
How AcquiredMethod of acquisition
LocationLocation in one cell (Room / Case / Shelf)
DimensionsDimensions in one cell (30 x 20 cm)

The first import reports: 4 objects added, 1 already in the catalogue, 1 added with a number another object has, 2 accessions added, 1 row skipped. It made accessions 2019.4, 2019.5 from the object numbers, and skipped: Row 5: object number "Box 7" is not year.accession.object, and there is no accession number column to file it under. Importing the same file again reports: 0 objects added, 5 already in the catalogue, 1 row skipped.

A cell starting with an equals sign, such as =HYPERLINK("x"), is exported as '=HYPERLINK("x") so a spreadsheet shows it as text, and the apostrophe is taken off again on import.

Common mistakes

  • Importing the original spreadsheet instead of a copy you have tidied.
  • Leaving the date order on the default when the sheet was typed day first.
  • Keeping several objects in one row, such as 12 plates in a single line.
  • Fixing numbers in the catalogue but not on the objects, or the other way round.
  • Deleting the spreadsheet after the import; keep it with the accession files.

Do it in the catalogue

Open the catalogue to number an accession and its objects, add photos and locations, and print the register, labels and receipts. Everything stays in this browser; nothing is uploaded and there is no account. Save the catalogue file after every session.

Frequently asked questions

My spreadsheet has one row per accession, not per object. Can I import it?

Split it first: one row per object with the object number. An accession with a single object can use its accession number followed by .1.

Will importing twice make duplicates?

No. A row whose object number (by value), object name and description all match an object already in the catalogue is counted as already there and left alone, and the report lists those rows by number. A row that shares only the number is imported as a different object and listed under Numbers, so a legacy duplicate number is never silently dropped.

Can I move to a bigger system later?

Yes. Export the objects and the accession register as CSV; most collection systems import CSV with a column mapper of their own.

Sources

Last reviewed 28 September 2026 against:

  • Cataloguing, the Spectrum standard, Collections Trust, read 28 September 2026. Used for: Catalogue records linked to objects by their unique numbers, consistent names, dates and places, and regular backups.
  • Numbering, Collections Trust, guidance updated August 2024, read 28 September 2026. Used for: Year based accession numbers with a running sequence and a further part for each object, full four digit years first, full stops as separators, no leading zeros, entry and temporary numbers attached reversibly, and renumbering discouraged.
  • Module 26: Numbering Systems, Oklahoma Museums Association, read 28 September 2026. Used for: Keeping a consistent existing system, the three number system for a new one, letter suffixes only for pairs and pieces that fit together, and separate numbers for the pieces of a set.

This guide is our own summary, not a rule book. Rules are paraphrased, never quoted at length, and if this page disagrees with the current published rules, the rules win. Curatorbook is not affiliated with, endorsed by, or connected to any museum, association, standards body or software vendor, including the Collections Trust (Spectrum), the American Alliance of Museums, the American Association for State and Local History, the National Park Service, PastPerfect, CatalogIt, MUSARCH, CollectiveAccess and Omeka.