Fill Empty Cells

Reports often write a department or date once and leave the following records blank. Fill Empty Cells copies an available value within a defined group or inserts an explicit fixed value. It is useful when the blank means that a known value continues. It does not predict an unknown observation, estimate a missing amount, or infer that two customers should share an address. The source of every copied fill remains available for inspection.

Choose whether the result should update the selected columns or add separate filled columns. Updating changes only the result copy; the imported table remains in the workflow history. Adding columns keeps the before and after values together for a direct comparison. Each selected column searches independently for its own usable source. If the department is known on one row and the date on another, one source record is not incorrectly reused for every field.

Choose files or Paste a table, then use Confirm file parsing to establish the header, delimiter, and logical records. Parse and preview shows an import sample; Confirm import validates the entire input. All values enter as text unless an explicit operation changes them. A missing cell in a short record is different internally from an empty string in a present field. Short records require explicit parsing consent before they are padded with missing cells.

Select Columns to fill. Open Empty definition if the defaults do not match your source. Missing cells and empty strings are selected initially; pure whitespace is not. You can add exact markers such as a particular dash or source-specific word. Zero and false are valid text by default. NULL and N/A also remain literal values unless deliberately added as markers. A marker considered empty in a target field can therefore be replaced, while the same original text remains available in the fill report.

Do not use a broad empty definition merely to make a report look complete. A zero sales amount and a missing sales amount usually express different facts. A space can be an accidental export artifact, but it can also distinguish a source convention. Compare a few source cells through the grid's detail dialog before changing the definition. The operation does not trim valid donor values, so a copied department retains the donor's original spacing and case.

Down searches earlier records within a group and fills later empty cells from the most recent original valid value. Up searches in the reverse direction and uses the next original valid value in the group's current order. Fixed value writes the text you supply into eligible empty cells. No mode automatically sorts the data. If the intended meaning depends on a date or sequence, apply an explicit sorting step before filling and keep that step in the workflow history.

A blank at the start of a downward group has no earlier donor and stays blank. A blank at the end of an upward group has no later donor and also stays blank. These are unresolved cells, not failed predictions. If a later valid department exists, upward filling may be appropriate, but it is a different business rule. Choose a direction based on the source convention rather than repeatedly trying modes until every blank disappears.

With Fixed value, the supplied text is the source explanation instead of another record. An empty fixed value is allowed, although it may leave the visible contents unchanged. The fill report still documents the eligible operation. Fixed filling is useful for a clearly defined testing label or a category explicitly agreed for missing values. It should not make unsupported claims about unknown data merely because a receiving application requires a nonempty field.

Select one or more Group columns to prevent values crossing a customer, project, or other boundary. The default grouping mode treats the same composite key as one group, even when its records are separated by records from other groups. Composite keys keep their components distinct; a delimiter inside one key component does not accidentally collide with a different combination. The tool preserves the current relative order of records within each group.

Separate contiguous segments treats each uninterrupted run as a distinct group. With keys A, B, A, the final A belongs to a new segment and cannot inherit a value from the first A. Under the default same-key mode, those A records are in one group and propagation is allowed. Both behaviors are useful, but they express different interpretations of a report. The grouping control states the difference so that separated sections are not silently combined.

Records with missing grouping values do not propagate to each other. This avoids treating every unidentified customer as one shared customer. If no group columns are selected, the entire file would become one group, so you must explicitly confirm that choice. That confirmation is part of the rule. It is usually preferable to select a genuine business key when filling customer-specific or account-specific fields.

Maximum fill run controls how many consecutive empty positions may be filled after an original valid value, or before one in upward mode. Zero means no run limit within the other configured guards. A limit of one fills only the first eligible blank after the donor. The newly filled cell does not become a fresh donor that restarts the count. Otherwise a limit intended to protect a long unknown interval would have no effect.

Consider group A with East, blank, blank. Down with no run limit yields East in all three records. A maximum run of one yields East, East, blank. The last position is counted as skipped by the run limit and remains unresolved. When another original nonempty value occurs later, the counter resets from that actual source. Each target column maintains its own donor and counter, so gaps in one column do not limit another column's independent propagation.

Use a table with group A containing 华东, blank, blank and group B containing blank, 华南, blank. Select the value field, group by the group field, choose Down, and run. All three A records contain 华东. The first B remains empty, while the last B receives 华南. The output record count is unchanged. A fill report connects each changed target to its donor record, and Unresolved fills explains the position that had no source.

Switch to a maximum run of one and repeat. The last A now stays empty because the limit is measured from its original donor. Try the nonadjacent-key sample separately to compare the two grouping modes. Preview sample and estimate helps demonstrate settings on early records, but a sample may omit a donor that occurs later. Use Run and review for a full conclusion, especially for Up and for groups spread across the file.

Export new copy can download the filled result, individual fill events, unresolved cells, or the complete report bundle. The shared exporter quotes separators and line breaks correctly and parses its output again before enabling download. Formula-prefix protection can change a copied donor or fixed value if it begins like a spreadsheet formula. Its impact is reported explicitly. Keeping original text is a separate choice, and the receiving application's type inference still needs attention for identifiers with leading zeros.

Save rules records field bindings, group mode, direction, empty definition, and the run limit. A new source structure must be reviewed before reusing it. Apply to workflow preserves an ordered step that can be undone or recomputed. A rule change makes the old result stale until another full run succeeds. Large fill audits have their own row and cell guards because several target columns can create more audit entries than input records. A blocked or cancelled run preserves the source; reduce the scope rather than assuming an incomplete audit is a complete result.