HOME > Chowhound > Site Talk >

Discussion

"Mark as duplicate" not working. Still shows in search results and causes posting error

I marked two Slanted Door records as duplicates yesterday. They are still showing up today in the search list with no indication they are duplicates

http://search.chow.com/search?query=S...

What does happen when you click on the link is that it redirects to the record marked as the original post.

Which would be fine except I made an error with one which I explain here
http://chowhound.chow.com/topics/6558...

So the first item in the list is directed to a firewood store.

Even WORSE ... this is really, really bad ... all three Slanted Door show up in posts and because I made an error on the first, everyone will be linking to a firewood store instead of this immensely popular restaurant.

Take a look at this
http://chowhound.chow.com/topics/6562...

  1. Click to Upload a photo (10 MB limit)
Delete
  1. To avoid errors, the Mark as Duplicate feature should confirm by showing the name of the places record rather than just accepting the number with no feedback.

    1. "Mark as Duplicate" did work for me, though there may be some latency or quirks.

      I marked 559125 (4 discussions) as a duplicate of 38 (64 discussions).

      Both links still show up in the search, but 559125 redirects to 38.

      One odd thing is that while the search results show 4 discussions for 559125 and 64 for 38, "Dig Deeper" for the consolidated record shows 551. I presume that has to do with the way the system now lumps linked topics in with keyword hits.

      1. Thanks for the detailed report on this.
        1. the duplicate record doesn't disappear immediately from search, but it should be gone within 8 hours, and in the meantime it will redirect to the correct record.
        2. We'll fix that slanted door/firewood link error
        3. We are likely going to remove access to the 'mark as duplicate' tool for 3 weeks until we launch an improved UI that will make it much less error-prone.

        28 Replies
        1. re: CHOW HQ

          It appears "Mark as Duplicate" has already been deactivated. So here's one for the Chow staff to fix:

          http://www.chow.com/restaurants/13968... is a duplicate of http://www.chow.com/restaurants/19653..., introduced in the redesign.

            1. re: squid kun

              Another:

              http://www.chow.com/restaurants/12642... is a duplicate of http://www.chow.com/restaurants/11788...

              I'm not looking for these duplicates; I'm just running across them randomly when I search for venues. They're cropping up in a majority of searches.

              They weren't introduced by volunteer users, so volunteer users shouldn't have to clean them up. They were introduced by the outside data provider brought in by Chow, so it falls to paid Chow staff to take care of them. This thread seems as convenient a place as any to flag dupes, so I'll continue to do so here, whether or not "mark as duplicate" is restored.

                1. re: squid kun

                  Another:

                  http://www.chow.com/restaurants/97074... is a duplicate of http://www.chow.com/restaurants/14612...

                  BTW according to the data provider's entry, the primary cuisine at this Sichuan restaurant is "Mexican."

                    1. re: squid kun

                      More dupes to be cleaned up ...

                      http://www.chow.com/restaurants/75114... is a duplicate of http://www.chow.com/restaurants/25395.... Note that the dupe - like many records supplied by Chow's data provider - appends automated-listings verbiage ("Ste Frnt") to the street address. The result is a bad map that places this Upper East Side restaurant somewhere on the Brooklyn waterfront.

                      http://www.chow.com/restaurants/11335... is a duplicate of http://www.chow.com/restaurants/51413.... Same map problem as with Nino's: the dupe appends extraneous words ("Frnt 1") to the street address, and the resulting map is wrong, placing this Manhattan restaurant in Brooklyn. The original record, whose address doesn't include "Frnt 1," has an accurate map.

                      http://www.chow.com/restaurants/13854... is a duplicate of http://www.chow.com/restaurants/38297.... Same map problem on the dupe.

                      http://www.chow.com/restaurants/13407... is a duplicate of http://www.chow.com/restaurants/6609/....

                      http://www.chow.com/restaurants/85843... and http://www.chow.com/restaurants/85843... are duplicates of http://www.chow.com/restaurants/85842....

                      http://www.chow.com/restaurants/12688... is a duplicate of http://www.chow.com/restaurants/19468....

                      http://www.chow.com/restaurants/11106... is a duplicate of http://www.chow.com/restaurants/15989....

                      http://www.chow.com/restaurants/10401... is a duplicate of http://www.chow.com/restaurants/16614....

                      http://www.chow.com/restaurants/96982... is a duplicate of http://www.chow.com/restaurants/26471....

                      http://www.chow.com/restaurants/13928... is a duplicate of http://www.chow.com/restaurants/27058....

                      http://www.chow.com/restaurants/13418... is a duplicate of http://www.chow.com/restaurants/50846....

                      http://www.chow.com/restaurants/10290... is a duplicate of http://www.chow.com/restaurants/14410....

                      1. re: squid kun

                        I'm running across a whole bunch for the SF area as well, and that's just random use for my own daily needs, not a campaign to find them. Not only are there duplicates of what was already in the Places database, there are duplicate records for a single restaurant that were supplied by the data provider and some refund would seem to be in order if they're priced by unique record. It seems that the 750K database of restaurants was loaded without a thorough data scrub to look for duplicates.

                        I've not worked with a database that large before, but I have purchased large contact files. I always directed IT to scrub for duplicates before loading it into our systems. The incentive was not only to have a clean database with all unique records, the payment was based on new records we didn't already have so I would be sure to kick out all dupes.

                        Given the magnitude of the problem that we're finding here, I would suggest the same to Chow.com with the caveat that I have no background on how this deal is set up. Trying to find them one by one, as you're doing, is not going to work.

                        If it's too costly to scrub the entire database (750K), I would suggest focusing on the ones with state code of NY and CA. Those two states had the highest number of existing records before the cutover and have the highest likelihood of duplicates.

                        Suggested data passes would be 1) match on name and address. After that is done, the remaining records would be scrubbed for 2) match on phone number and address. And then, what's not kicked out in either of those two would be scrubbed for 3) match on address. Then a paid staff member would need to review the reports, decide what's valid, and work through removing the duplicates.

                        1. re: Melanie Wong

                          Thanks for the redirect from the other "New Restaurants Section" thread. I agree with Melanie--any of us reporting one-off duplicates here on this thread isn't a viable option. The sheer volume is too great. So while I was going to post some of the issues from the New England board (specifically CT), instead I'll wait to hear as to how Chow.com plans to proceed. Help me help you...what can we do to assist, if anything?

                          ALSO, I've been having fun updating local spots on my board. But as one of my friends pointed out, it's too bad there is no "flag" that sends an update to anyone's reading list. I know there are people I'm reading who are updating places, adding photos and doing "quick reviews." I'd love that info to land in "My Chow." That would be a really nice enhancement for a future wish list.

                          1. re: Melanie Wong

                            This is absolutely happening quite a bit.

                            What is really difficult is that opentable link.

                            Also ... please, please, please ... on my knees please ... bring back the section that had explicitly linked records.

                            Yes, I realise there are places in the country not as active as the metro areas, but this is throwing out relevant info for the largest segment of the board for the few. There were two sections before. Please reinstate that.

                            Here's an example, because I truly don't know what to do here

                            There was a great report on Crab House on Pier 39. It was buried in a post asking for seafood recommendations ... I think

                            If we could find explicitly link records, I could find that report which would be useful to another poster.

                            It would also clear up which of the three restaurant records were duplicates.

                            I am guessing this is the original because it has a lot of info
                            http://www.chow.com/restaurants/5944/...

                            This is probably a duplicate, though it is possible the report I want is linked to it.
                            http://www.chow.com/restaurants/64661...

                            This is the record generated by the load because it has the opentable link
                            http://www.chow.com/restaurants/65430...

                            I was just going to mark the last two as duplicates in the title, but I am unable to move that opentable info into the real original record that possibly has all the links

                            Even though those links can't be found without literally spending at least a half hour to open and read every single post currently listed, I am hoping that those links will someday be useful again.

                            1. re: rworange

                              I'm confused. CHOW HQ above said they were disabling "Mark as duplicate"--can we still use this feature? I will if I can.

                              1. re: kattyeyes

                                Nope. Until this was fixed I just going to put the word "duplicate" in the title for the time being so the correct record would get linked

                                I know squid kun is putting duplicates in this link, but is the Chow staff going to do anything about these? Or will they just wait until the updated interface is installed and then posters will need to take care of it.

                                1. re: rworange

                                  OK, I guess I'll wait to hear next steps. Thanks for clarifying. You're in a far bigger area than I am. Once I know how to proceed with duplicates in my area, I won't have as long as list as you or squid kun to deal with.

                                  Any intel on "Quick Reviews" and where they "live" in the land of Chow? It's disconnected that they appear to exist outside Chowhound (and just on Chow)...guess that's why they don't link to your profile or anything. Bummer. I like the flexibility of being able to post pics with captions, but then the posts are out there in "no man's land" of CH.

                                  1. re: kattyeyes

                                    When you post on Chowhound, clicking on the review box puts your report in the Quick Reviews section of the restaurant record. The advantage is you have a two hour window to edit. If you enter your review on the restaurant page it cannot be edited.

                                    1. re: rworange

                                      Too bad that relationship isn't a two-way street. It would keep the boards current to have updates on a restaurant record trigger something new to read on the local board. Oh, well. Maybe if anyone finds that idea of value, it can be a future implementation. And, yes, I noticed the no edits part. I will have to remember that for future updates.

                                  2. re: rworange

                                    I cleaned up all the ones above, I think.

                                    1. re: Jacquilynne

                                      Help me understand--would you like us to post duplicates here for cleanup, or do you have some other solution in the works? I'm happy to post a few. My list won't be long.

                                      1. re: kattyeyes

                                        If you've found them, you may as well post them, and I'll fix them. I wouldn't go off on a witch hunt for duplicates, though, as it seems like it's something that's going to need a larger solution than them being fixed one at a time.

                                        1. re: Jacquilynne

                                          Thanks, please see this post and the two that follow it (one is rbailin's):
                                          http://chowhound.chow.com/topics/6558...

                                          1. re: kattyeyes

                                            Talking about this with Engineering they did ask that people not make edits to the names of duplicates that you find, clean them out, etc. Until Mark as Duplicate is returned to contributors, or Engineering is able to find a broader solution, please just report any duplicates you find here, and I'll fix them.

                                              1. re: Jacquilynne

                                                http://www.chow.com/restaurants/10626... is a duplicate of http://www.chow.com/restaurants/19610.... The address on the dupe includes the robo-listings language favored by our outside data provider (not just "342 E 6th St" but "342 E 6th St Frnt A"). Perhaps not coincidentally the map on the dupe record is way off, placing this East Village restaurant in Park Slope, Brooklyn. (The map on the original is correct.

                                                )

                                                More duplicates ...

                                                http://www.chow.com/restaurants/12864... is a duplicate of http://www.chow.com/restaurants/13716....

                                                http://www.chow.com/restaurants/13557... is a duplicate of http://www.chow.com/restaurants/38542....

                                                http://www.chow.com/restaurants/92977... is a duplicate of http://www.chow.com/restaurants/9453/....

                                                1. re: Jacquilynne

                                                  Again, as squid kun points out, the duplicate problem for records supplied by your data provider should not be the responsibility of users to clean up. He and others have already been down that road with the old system.

                                                    1. re: Jacquilynne

                                                      http://www.chow.com/restaurants/12697... appears to be a duplicate of http://www.chow.com/restaurants/56866....

                                                      I say "appears to be" because the later record bears the same address and phone as the earlier one but a different name - probably one that appears on a business license, but not one recognizable by the dining public. This is another result of buying information from an outside data provider and putting it on the site without sufficient checking.

                                        2. re: rworange

                                          It looks to me like the url numbers of the original Places records were retained. Before the cutover, I think they went up to about 60000 or so (squid kun, do you recall?). So, anything larger than that is new from the data service.

                              2. For Euro Pane Bakery in Pasadena, CA:

                                http://www.chow.com/restaurants/46538...

                                is a duplicate of

                                http://www.chow.com/restaurants/3858/...

                                (The latter spells Euro Pane correctly as two words.)

                                        1. duplicates:

                                          ABC Sea Food in Foster City
                                          Acme Chophouse in SF
                                          Acme Express in Martinez
                                          Ravenous Cafe in Healdsburg

                                          5 Replies
                                            1. re: Melanie Wong

                                              Yoshi's branches in San Francisco and Oakland both have duplicates.

                                              I noticed that the records I reported in the post directly above are still in the system. Does it take that long for the consolidate to update?

                                              1. re: Melanie Wong

                                                No, it's pretty much instantaneous. I, however, am not. I just hadn't done them. I have now, though I can't find a duplicate for Acme Chophouse.

                                                  1. re: Jacquilynne

                                                    The duplicate and original for ABC Sea Food are still in the database. Someone just linked to the duplicate in a posting.

                                                    ETA: I marked it as a duplicate using the new interface. Let's see if that sticks.

                                              2. Three listings for the newly reopened Mazzarino's Italian Restaurant in Sherman Oaks, CA.

                                                The correct listing (12920 Riverside Drive):
                                                http://www.chow.com/restaurants/7295

                                                Incorrect addresses:
                                                http://www.chow.com/restaurants/30435
                                                http://www.chow.com/restaurants/648499

                                                          1. Nation's Giant Hamburgers - 13296 San Pablo
                                                            http://www.chow.com/restaurants/15901...
                                                            Duplicate
                                                            http://www.chow.com/restaurants/84572...

                                                            Actually it looks like all the Nation's records were duplicated. Since I added them all a while back and made the name consistant, all of Nation's Giant Hamburgers are originals
                                                            http://search.chow.com/search?locatio...

                                                            While the duplicates are "Nation's Giant Hamburgers & Great Pies"
                                                            http://search.chow.com/search?query=%...
                                                            While

                                                                1. Oliveto Downstairs & Cafe appears to have been inappropriately merged with Oliveto Restaurant. Same address, different menus, prices, hours, and reservation policies.

                                                                  http://www.chow.com/restaurants/27015...

                                                                  1. I can't mark this one as a duplicate because the restaurant name doesn't appear on the list:

                                                                    http://www.chow.com/restaurants/84756...

                                                                    Original: http://www.chow.com/restaurants/34/two

                                                                    1 Reply
                                                                    1. re: Robert Lauriston

                                                                      There are three Chambars in the resto listings for Western Canada -- only the third has discussions linked to it.

                                                                    2. Two entries for kuletos in Burlingame and the resto is closed.

                                                                      1. Two listings for Langer's Delicatessen in Los Angeles:

                                                                        Correct:

                                                                        http://www.chow.com/restaurants/2287/...

                                                                        Incorrect:

                                                                        http://www.chow.com/restaurants/63005...

                                                                        The "Mark As Duplicate" button still doesn't work ...

                                                                        1. Romeo's in Petaluma has two records. Blue Mermaid in SF has a duplicate.

                                                                          1. When I couldn't pull up a place I wanted to post about in the linking function, I went to look for it inR&B search. Not there anymore when I searched by name. So I pulled up everything in Salinas, CA. There are a bunch of duplicates created by the data provider. I'd entered a lot of places in my hometown that didn't get recognized in a purging of duplicates.

                                                                            So what's the plan for getting rid of the duplicates in those areas where users had populated the database in the early days of Places? Bear in mind that NY, LA, and SF (I wouldn't consider Salinas vital) are the primary battlegrounds for going head to head with Yelp. This feature dumped a bunch of bad entries in your key competitive markets.

                                                                            3 Replies
                                                                            1. re: Melanie Wong

                                                                              Engineering does have something they're working on for finding possible duplicates so we can fix them.

                                                                              1. re: Jacquilynne

                                                                                Thanx J. I'm mostly using the site on mobile and it's just too painful to deal with the dupes I run across inthat format. Glad to hear abt this.

                                                                                1. re: Jacquilynne

                                                                                  I wanted to check on the status of a fix for scrubbing out duplicates. I noticed that several new records, all with Opentable links, were loaded into the SF Bay Area database today. With just a cursory look at the ones I thought were likely and not checking all of them, I knocked off about a dozen dupes in about 50 new records. Some of these, such as Mezza Luna in Princeton, were an exact match on name and street address, so if the duplicate checking program is in place, it's not working. More attention should also be paid to "new" records feeding in from Opentable, as many are not new and create a duplicate of an existing place.

                                                                              2. Boccalone Salumeria in SF had two duplicates from the data provider. I've marked them but wanted to note it here that you're getting doubles from your vendor.

                                                                                1. This entry for Two:

                                                                                  http://www.chow.com/restaurants/847564

                                                                                  Is a duplicate of:

                                                                                  http://chowhound.chow.com/restaurants/34

                                                                                  However, I can't mark it, because when I enter "two" in the field, I get a bunch of places with names starting with "2" (many of which are out of the area).

                                                                                  2 Replies
                                                                                  1. re: Robert Lauriston

                                                                                    Yeah, I've been piling up some examples of the same problem. The drop-down doesn't go beyond around 10 venues, and at least on my browser it can't be scrolled past that point.

                                                                                    1. re: Robert Lauriston

                                                                                      Actually, *all* of the restaurants on the list were out of the area.

                                                                                      Worked OK today, so maybe it's fixed.

                                                                                    2. http://www.chow.com/restaurants/99320...

                                                                                      Can't flag this as a dupe; the original record - http://www.chow.com/restaurants/18573... - doesn't show up in the "Mark as Duplicate" drop-down list.

                                                                                      4 Replies
                                                                                      1. re: squid kun

                                                                                        I've found this with several places that are beyond the poorly selected 60-mile radius from SF that limits place linking. That is, many places discussed on the SF board fall in the 60 to 95-mile range and seem to be assigned to a different region in the r&b database. It's easier to mark dupes now that the drop-down list is limited to region and not the whole database, but things like this fall in the cracks.

                                                                                        1. re: Melanie Wong

                                                                                          Melanie,

                                                                                          I don't think it is a radius thing here. Besides that, there are a few other things that might cause the problem, so I tried the duplicate option on squid kun's record. It wasn't the usual stuff. However, what I was getting, was the same record in the drop down last as the record to be marked that is for

                                                                                          http://www.chow.com/restaurants/99320...

                                                                                          that record was in the drop down list.

                                                                                          For the other record, it was the other record.

                                                                                          Something changed last night. I've been doing some lists and suddenly not only the name was in the drop down list for ... lists ... but also the address was carried into the box. There were some other weird things that just switched suddently.

                                                                                          1. re: rworange

                                                                                            I just tried it again. I'm not getting either record for an exact name match, Han Shin Pocha, in the drop-down menu, trying to access "mark as duplicate" from each one.

                                                                                            1. re: Melanie Wong

                                                                                              rworange and Melanie - thanks to both of you!

                                                                                              So I'm not going crazy. OK, I might be, but not because of this.