Description:
When a table of data has auto filter set on all the columns and then a filter is made which will "capture" say three rows of data and then all the cells in a column of the filtered rows are selected by cursor, then the automatic summary of "rows" "total" and "average" presented at the bottom of the sheet does not reflect reality. Furthermore, the line numbers on the left of the sheet change with no apparent connection to either the "true" line number in the sheet or the order in whioh the rows have been sorted.
In some instances, the summary identifies the number of rows "captured" as the number of rows that WOULD HAVE been captured if the sheet had not been filtered to exclude the majority of those rows. In other words, it appears to be the count of actual rows in the sheet commencing with the first pertinent row and culminating on the last pertinent row.
However, even this inconsistency is not consistent. If the now current three rows are sorted by using the "sort ascending" feature of the filter dropdown at the top of any column other than that which has been filtered and then the all the cells in any column are selected the count of rows again changes. But there are still only three rows selected.
It only presents the correct number of rows if the column which has been "filtered" is then "sorted ascending". There is no visible difference following this "sorting" but the summary correctly counts only the three rows present.



Steps to Reproduce:
1. Open the spreadsheet attached
2. Select only "red cross" with the "type" filter
3. Select the three cells of the "cost" column
4. Note how the calculated selection summary is inconsistent with reality
5. Observe the sheet line reference numbers at the left margin of the sheet
6. Select "sort ascending" from the "cost" filter
7. Observe the calculated selection summary has changed but is still inconsistent with reality
8. Observe the impact upon the line reference numbers at the left margin of the sheet
9. Select "sort ascending" from the any other filter
10.Observe the calculated selection summary has changed but is still inconsistent with reality
11.Observe the impact upon the line reference numbers at the left margin of the sheet
12.Select "sort ascending" from the "type" filter
13.Observe the calculated selection summary has changed and now reflects reality
14.Observe the impact upon the line reference numbers at the left margin of the sheet
15.Select any random elements desired and observe the effect upon the calculated selection summary and the line reference numbers at the left margin of the sheet
16.Release the selection filter on "Type" by selecting "all" and observe that all the lines have been sorted by whatever was the last sort. Clearly, this has not been taken into consideration when calculating the summary and probably also not correctly handled in the closed printing error bug report at 17.
17.Refer to bug 117600 - closed by Joel Madero in a fit of pique :) - and convince yourself these are not related issues
In total, four attachments will be added. The calc file identified together with screen images highlighting the anomalies.

Actual Results:
Incorrect summaries (incidental to the printing error report 117600)

Expected Results:
Correct summaries and correctly formatted printing


Reproducible: Always


User Profile Reset: No



Additional Info:
[Information automatically included from LibreOffice]
Locale: en-GB
Module: SpreadsheetDocument
[Information guessed from browser]
OS: Windows (All)
OS is 64bit: no