Review and fill missing values without inventing observations

Quick answer

Use Data Cleaning → Fill Missing Values only after deciding what each blank means; preserve measurements that are genuinely unknown.

Before you start

Zoho Sheet web editor in an English desktop browser; rolling service, official documentation reviewed 20 October 2026. Not desktop-app or mobile instructions. Use an editable practice file containing only synthetic data. Keep the original unchanged and record the file name and storage location so that you reopen the same copy. These are documentation-based instructions and independently checked examples, not a claim of native application testing.

Prepare an auditable fixture

Create these headers in A1:D1: Record, Team, Hours, Review. Enter five records:

Row	Record	Team	Hours	Review
2	R01	North	4	Checked
3	R02	North
4	R03	South	0	Checked
5	R04	South	6
6	R05	North		Checked

Leave the displayed blank cells genuinely empty, rather than entering spaces. Keep an untouched copy of these rows. The known hours total 10, but that is not a complete total for all five records. A zero in C4 means an observed zero; a blank in C3 or C6 means the observation has not been supplied.

The safe task is to label the missing review state, not to fabricate hours. Our explicit business rule is: an empty Review means Needs review. Record that rule beside the table, along with the date and the person approving it.

Fill only the justified field

1. Open Data → Data Cleaning → Fill Missing Values.
2. In the data-cleaning panel, choose the Review column. Check that the detected empty cells belong to D3 and D5 in the practice data, not the Hours column.
3. Choose the option that supplies a specific value and enter Needs review.
4. Review the affected cells and apply the operation using the panel’s displayed confirmation. If it proposes more than the two intended records, cancel and correct the selected scope before proceeding.
5. Compare the result against the untouched copy. R02 and R04 should now say Needs review. R01, R03 and R05 must remain Checked; the numeric column must be unchanged.

The documented alternatives are filling from the previous cell or using the most frequent value in the column. Neither is a substitute for knowing the meaning of a blank. Carrying North forward into an unknown Team could assign a record incorrectly. Filling missing Hours with the most frequent observation would disguise an assumption as a measurement.

Verify and challenge the rule

Count two Needs review records and three Checked records. Confirm C4 is still 0 and C3/C6 are still empty. Now enter 5 in C3 because a genuine observation has arrived. The known-hours total becomes 15; one record remains unknown. Nothing about the review-state fill should prevent this later correction.

If a source file uses an empty string formula or whitespace instead of an empty cell, inspect it separately. The source describes empty-cell detection, not every possible blank-looking representation. Do not promise that one cleaning operation catches all of them.

Recover and report honestly

If you filled the wrong column, undo immediately and compare to the original, or return to the preserved copy before continuing. Never repair a mistakenly imputed numeric column by guessing which zeros were genuine.

When sharing the cleaned result, retain a short change log: two review states filled under the stated rule, no missing hours imputed. Reopen the saved file and inspect those exact records. Exporting the cleaned values does not carry the original blanks or your rationale unless you include them deliberately.

Related tutorial

Continue with standardize misspelled labels data cleaning.

Official sources

Documentation reviewed 20 October 2026.

• Fill missing values