Issue Number | 3859 |
---|---|
Summary | Table display problems in PP and on Cancer.gov |
Created | 2015-01-15 17:12:19 |
Issue Type | Bug |
Submitted By | Osei-Poku, William (NIH/NCI) [C] |
Assigned To | Englisch, Volker (NIH/NCI) [C] |
Status | Closed |
Resolved | 2015-01-21 17:42:06 |
Resolution | Fixed |
Path | /home/bkline/backups/jira/ocecdr/issue.145016 |
It looks since the last WCMS release some tables which hitherto displayed correctly (in PP and Cancer.gov) are now displaying incorrectly (in PP and Cancer.gov) . They are usually misaligned with some of the text in the last columns creating another incomplete column to the right of the table. The tables mostly affected are ones which have split rows and columns to the right hand side of the table. Meanwhile the tables display correctly in XMetal. Volker took a look and thought that it was a bug in the XSLT transformations of the tables so he created a Jira ticket for the WCMS team. Here are two examples.
http://www.cancer.gov/cancertopics/pdq/treatment/childependymoma/HealthProfessional/page4
http://www.cancer.gov/cancertopics/pdq/treatment/child-astrocytomas/HealthProfessional/page3
The issue has been reported as OCEPROJECT-1960.
The problem of processing the tables shows up when there is a cell spanning across several rows while there are also cells spanning across several columns within the same "row-spanning" block of rows such as this:
|--------------------|
| | a | b |
| lala |-----------|
| | c |
|--------------------|
In this situation the "converter" thinks there are a few cells
missing on the row with cell c and adds a couple of empty cells
as if the rowspan did not exist.
I removed the piece of code that's responsible for adding empty table
cells because I don't believe XMetaL would produce tables with implicit
(empty) table cells.
This needs lots of testing.
You can use this link on DEV in order to test the table
changes:
https://cdr.dev.cancer.gov/cgi-bin/cdr/show-document_work.py?doc-id=614165
(change the doc-id for testing different summaries).
Verified on DEV. Thank you!!
William, I'm hoping you've not just looked at the tables that displayed the error but also at other complicated tables to ensure I didn't mess up anything with this change. I don't think I did but I would feel better if we looked at a few tables that worked fine as well.
Yes, we looked at other tables within the same summary and they all looked good. However, if you want me to do more testing, I will be happy to do that.
Here is the list of affected tables:
CDR614165: Ch Astrocytoma, Table 4: http://www.cancer.gov/cancertopics/pdq/treatment/child-astrocytomas/HealthProfessional/page3
CDR62843: Ch Ependymoma, Table 1: http://www.cancer.gov/cancertopics/pdq/treatment/childependymoma/HealthProfessional/page4
CDR62846: Retinoblastoma, Table 7: http://www.cancer.gov/cancertopics/pdq/treatment/retinoblastoma/HealthProfessional/page3
CDR62846: Retinoblastoma, Table 8: http://www.cancer.gov/cancertopics/pdq/treatment/retinoblastoma/HealthProfessional/page4
CDR62741: Adenocarcinoma, Table 4, http://www.cancer.gov/cancertopics/pdq/treatment/esophageal/HealthProfessional/page3
CDR62917: Melanoma, Table 8, http://www.cancer.gov/cancertopics/pdq/treatment/melanoma/HealthProfessional/page4
CDR62910: Prostate, Table 1, http://www.cancer.gov/cancertopics/pdq/treatment/prostate/HealthProfessional/page2
Is this a list of the table that currently have the problem on Cancer.gov for Margaret's information or is that a list of tables you've testing and these are now displaying incorrectly on DEV?
This list is a complete list of tables that have been misaligned due to the problem you identified. I have not tested them yet using the link you provided. You're right, they are only for Margaret's attention (For which I sent her an email as soon as I posted the list).
I'm glad! I didn't like the idea of having to fix another bug in the table templates. :-)
CDR62910: Prostate, Table 1
Please note that this table is easily corrected by removing one of the superfluous description columns.
Volker: do we know whether this is in some queue for WCMS to address the problem?
Yes, we know this from the first comment of this ticket. It's been fixed as part of DevonRex-IT1.
I checked the tables listed in this issue and they all appear to have
been fixed. I'm guessing the fix was part of the latest WCMS hot-fix
release.
Please verify on PROD.
Since all of the English summaries had been reprocessed last weekend the tables look fine, however, the Spanish summaries with the table issue would still need to be reprocessed in order to fix those.
Verified on Prod. Thank you!
Elapsed: 0:00:00.001695