bluearrow

Any ideas on this topic? The two main reasons why a change is needed are: - Near future prospects of a large number of concurrent online conections. - The need to provide faster reporting. We thought this could be better server from MySQL. Thanks

Hi, I have a large database in FM12. I need to provide scalability and future proof the development in the middle term, without loosing the fast adaptability and development of FM. I was thinking in using MySQL as back-end and keep using FileMaker as front-end, ensuring RAD when needed. Is there any way to do this without the need to reprogram or recreate the DB in MySQL and/or reprogram the FM DB? The idea is to find a simple path for this task, if there is any. Thanks

Latency, uhmm. I will check. Would a remote desktop connection to a local machine work better than FM own port connection? It does not sound logical, but...

Hi, Three remote offices connect to a hosted FM solution (larger than 50.000 records, many tables and files, separation model in place) using port 5003. The database is located in office 1. Office 2 is in the same country, different city, office 3 is in a different country. Both remote offices complain that the FM solution is slow when showing records and other operations (not even mention showing a report). They are using 4 Mb and 10 Mb symmetrical DSL lines. Locally the solution works fine. Besides FM, Internet connection in every office is also fine all time (no congestion or slow response). The database does not contain fancy images, icons, container fields or similar. Just plain text data and standard FM graphic features used with moderation. What bandwidth would you recommend? Bandwidth apart: Is this normal under this circumstances? What other options besides optimizing the database further? (this has been tried before with no luck) Thank you

Thank you very much for the file. Your solution works fine. I will take into account your comments about a cleaner solution (in fact, I am using FMPA 11, not 9; I have modified that on my profile). Thanks again

Hi, I have a table for items. Items are defined by a Class, Type and Color combination. Items have a price. For instance: Item1: 1, 3, 4, 4 (Class, Type, Color, Price). Item2: 2, 3, 6, 8 Item3: 2, 3, 6, 2 Item4: 1, 3, 4, 2 Item5: 2, 3, 6, 12 Item6: 2, 1, 1, 6 What I need to obtain is the average price for each Class, Type, Price combination, not in a report, but in a field within the table. The field should update every time a price change or a new combination is created. In the sample above it would be: 1, 3, 4 = (4 + 2) / 2 = 3 (average price for item with combination 1, 3, 4) 2, 3, 6 = (8 + 2 + 12) / 3 = 7,34 2, 1, 1 = 6 / 1 = 6 I do not mind creating a new table or using a custom function, if that is the best route to go. Note: I was trying to attach attach a file (tried FP7, xlsx, csv, htm) with a number of records and sample values, for your convenience, but for some reason I always get the message "You aren't permitted to upload this kind of file". If you know the reason and solution, I will upload the file. Thanks in advance

Clear enough. Thank you both. Regards

Hello Comment, You are right. In the first example, I was thinking in having different records, some records with teachers and classes ID (for instance) and other records with students and courses IDs (but without teachers ID). I could then have a JOIN file with different records joining different tables. The alternative would be to have different JOIN tables for each case. Does this make sense? Having only one JOIN table seems simpler, but I am not sure it is good practice or if could have negative implications of which I am not aware at the moment. Sorry for not being clear in the examples. Regards, Robert

Hello, From a design point of view, in a set of related tables within a database: is it recommended or optymal to have multiple join files or a single one. For instance, given: teachers students classes teachers_exams courses We could have: teachers JOIN1 classes students JOIN1 classes teachers JOIN1 teachers_exams students JOIN1 courses teachers JOIN1 courses or teachers JOIN1 classes students JOIN1 classes teachers JOIN2 teachers_exams students JOIN2 courses teachers JOIN2 courses Thanks

Thanks Steve. Checked. The hosted files were fine, no auto log on credentials. However, the opener file did have the admin account as log on account. This explains the errors accessing the hosted file. Now I do not see any errors on the server log files. Could this be related to the connectivity problems when restarting DC servers? I do not see the relationship among them. Regards Roberto

Thank you Wim. I am also surprised, as we have previously used and installed Server 5, 6, 7, 8, and 9 on the same network without any errors like this (at least, not that we have seen them or users have complained about it). Yes. And the description states: "Client "Whateverusername" authentication failed on database "Databasename" using "Admin [fmapp]". Yess. Those entries exist and they correspond to "Failure audit" under Categories "Account Logon" and Account Logon". For Log/Logoff the message is: "Logon Failure: Reason: Unknown user name or bad password User Name: Admin Domain: SERVER9 Logon Type: 3 Logon Process: Advapi Authentication Package: Negotiate Workstation Name: SERVER9 Caller User Name: SERVER9$ Caller Domain: DOMAIN1 Caller Logon ID: (0x0,0x3E7) Caller Process ID: 1308 Transited Services: - Source Network Address: - Source Port: - and for Account Logon: Logon attempt by: MICROSOFT_AUTHENTICATION_PACKAGE_V1_0 Logon account: Admin Source Workstation: SERVER9 Error Code: 0xC0000064 The server where FMS is located is a member server named SERVER9. The domain name is DOMAIN1. I am not sure why it is using the Admin account for all authentication, neither why it calls SERVER9 a Domain (it is not, neither it has AD; it is part of Domain1, and the same goes for all clients). Yes. I have seen this and it makes it difficult to test if FMS and clients communicate with only one DC server. Also, I have noticed that there is a delay when users start Windows and when user may start the Filemaker database. This is normal under windows. This is not easy, as those DC serves also run other tasks that cannot be stopped for longer than 30 minutes. However, replication among DC servers is fine, and user accounts have been up and stable for years. The problem is not related to account creation or password change (we do change them twice a year, but the issue was present before and is present after this changes). Any ideas Thanks

Hi, I have a Windows 2003 server / XP workstation /Filemaker Server 10 network, single domain. All latest patches are regularly applied to servers, workstations and the FMS. FMS is intalled on a member server machine (no AD on that machine). There are two DCs on the network (two servers containing AD and replicating to each other). The network works fine in all respects and all software is also working fine. However, client Filemaker Pro 10 machines fail Single Sign On under the following conditions: - Upon restarting one of the DC servers (which does not make any sense, as FMS could use the other to reach AD; in fact, users can restart their machine and logon to the network without one of the DC servers, using the other to authenticate). Users who are logged on to the domain and to FM (Filemaker database is open) when this happens, do not loose their connection to the database. They can work as usual. The problem is for users who do not have FM database open, or when the user closes the database and tries to reopen it. Filemaker opens, but authentication fails. - Upon restarting the FMS machine without restarting the DCs (sometimes FM Single Sign On works after waiting several minutes - more than 15 minutes) Users are logged on to their machine as Users, not as Administrators. They have no problems to log to the domain, under any circumstances. We have tried with new Windows 7 workstations and the problem with Filemaker SSO is the same under the above conditions. This problem did not happen using FMS 9, on the same machines, same users, same AD FM groups, same database and identical SSO mechanism and opener file, on the same network. It seems to be an issue of FMS 10 (it happened with FMS 10 v1 and it also happens with the latest patch applied). I have seen some post with similar problems. Any ideas on how to solve it? It is really annoying to have to tell users to wait 20 minutes to use the database, or have to restart server several times because we are not sure what is going on. Thanks in advance

Hi, How would you manage counting the frequency of all values in a field within a table? For instance, there are 10 records in Table1. Field1 in Table1 has the following values: Empty, 2, 4, Empty, 3, 1, 4, 4, 1, 4 The information I need would be: Empty = 2 times 1 = 2 times 2 = 1 time 3 = 1 time 4 = 4 times I guess this can be handled by a self cartesian relationship but I do not want to create more than one relationship or several different calculated fields. Field1 values are not part of a list: they could be any numeric value, included zero or even the field could be empty). Thanks in advance

For some reason, I am getting an error while trying to download your attachments. It seems a site error, though: "You have requested a resource that does not exist. The above information has been logged." Anyway, are you talking about a scripted constrain? Because the users do not have access to any of FileMaker menus. Just a custom search dialog with search fields an a Find button. I can see it is possible to use the GTRR, but it seems to me very complex. I will have to include in the search script a conditional line in case the user request a holiday restriction. After that, go to the child table, then GTRR to return to the parent records, then omit, then constrain the search with the rest of the search request. I was thinking in something much, much simpler like treating this date range search like the rest of search fields, if possible at all (I am lazy and I like simple coding, I recognize it). Maybe your file contains that kind of solution. Initially, I thought of creating an "inverse" relationship between the parent and the child table based on the date fields (in the child table) and a new field in the parent table, but it did not work. Then I thought of storing all child ranges in a field in the parent table (not a good idea, I know) as carriage return separated numeric values, to avoid searching in the child table. Not sure how to best accomplish it, anyway...

Thank you Vaughan. That is indeed the second part of the problem: how to search for a date range using a date range (really, looking for records that do not match the searched date range). I do not have any idea how to do this. But the first part is how to include the date search into a more complex search, as described in my reply to Søren.

Hi Søren. Thanks for your reply. Sadly, I cannot user what you propose. Showing omitted will display omitted child records that comply with the date criteria, not parent records. This could be solved and show the related parent records, but: what about other search criteria? To give you a better picture: - The user is looking for Contacts. - The search is started from the Contacts table. - The date range is a holiday period - A contact might have several different holiday periods a year (or over the years). Obsolete periods are not deleted. - the user might request contacts whose name start with letter S, or contacts from a specific client, or contacts whose name start with letter S, from a specific client, and which are not currently taking holidays (this are just examples, the real search screen includes many more search fields). What I want to clarify with this is that the holiday search is not an isolated search (and cannot be), but part of a much more complex search. There are many related tables that could be queried. It is simply not possible to go to the related records and look for a date or a date range because there might be other criteria in the search and this criteria is not part (neither is present) in the Holiday table. It is the Contacts table the one that holds all relationship, not the Holiday table. Does it give you a better view of the problem?

Hi, Finding all related records within a date interval is easy, assuming you are searching inside a single date field in the related table. You just specify StartDate...EndDate. I need something more difficult: I have 2 related tables: 2 date fields in the child database (StartDate and EndDate). A one to many relationship. Every child record has a date interval. I need to find all parent records that, given a search date (single date )or interval (StartDate...EndDate), have no child records in which that date or interval is present. And I mean parent records: the search result must show parent records. All searches are scripted and carried out from the parent table. The user might select other search parameter that are not relevant to this issue but might be part of the search conditions. There is no way this can be changed (I cannot script just to search for this interval, for instance, going to the child table, searching into the dates fields and then returning to the parent table). It must be a direct search. An example: Child records for Parent record 1 Record 1: StartDate 03/March/2008 EndDate: 15/March/2008 Record 2: StartDate: 07/October/2008 EndDate: 09/October/2008 Child records for Parent record 2 Record 1: StartDate: 05/April/2008 EndDate: 16/April/2008 Record 2: StartDate: 07/October/2008 EndDate: 12/October/2008 Date searched: 08/October/2008 Expected search result: No records founds Date searched: 09/March/2008 Expected search result: Parent record 2 Please let me know if I have not explained clearly what I need. Thanks

Thank you all for your ideas and comments. All seems excellent advise and very practical too.

Thanks. That solved the display issue. But: would it not be faster for FileMaker to create the record in a child layout with no records at all, instead of having thousands of records and adding one new record at the end? Just wondering...

Hi, I do not allow users to create child records using a portal. Everything is scripted to have better control on what is going on, avoid user errors and have additional task performed during record creation. Thanks

Hi, For creating children records starting from the parent table, I use a script like: Go to layout (target table for child records) New record/request (new child record) (Capture the child record ID) (Do some other tasks) Go to original layout The problem is that the database is growing and going to the child table layout is not practical anymore because when it does, it shows all child records (several thousands) before creating the new record. I cannot use a Go to Related Record because there might be no related records. Scripting a Find before the going to the layout is not practical either (does not make any sense, anyway: what am I going to find?). How can I script new child record creation and avoid showing all child records in the child layout? Take into account that all this is scripted, the user does not have any control over the process except pressing the New child record button from the parent table. Thanks

Hi, Is there any way to create a list of all layouts that are not used in any script? Maybe a plugin or a script? I know that using a DDR or some external reporting product I could find this out, but one by one (not nearly practical when you have hundred of layouts and scripts). The reason I need this is to delete all developer garbage layouts. Thanks in advance

It worked! Thanks Comment. Strange it needed the layouts formatting though...

Hi, I have found something that seems to be a bug. A table has a numeric field. If I export records from that table as CSV, all records that have a value lower than 1 and higher than zero (any decimal number such as 0.32) do not export the leading zero. This is a big problem for exchanging data among applications. One important consideration is that I am using an english version of FMP Advanced v3 in a non english OS (Windows XP spanish), using standard regional options for spanish (decimal symbol is a comma, thousands separator is a dot). However, this should not negatively affect exports from FM or any other program, by the way. From what I have found: 1.- If I only export that numeric field, FM exports the zero (for instance: 0,32) 2.- If I export more than one field, FM adds Text qualifier - "" - and text delimiter (a comma). This might be the source of the problem, but if it is, is a bug. 3.- Exporting more than one field removes all zeroes in decimal numbers lower than 1 and higher than zero (for instance: 0,32 becomes ,32). 4.- This problem did not happen in versions prior to FP v7. FM exported the zeroes. The main difference I have noticed is that in older versions, FM did not use text qualifiers. Of course, it is much better to have text qualifiers, specially while exchanging data with Excel and similars. 5.- The problem seems to happen also with other file format (I have tried some, not all of them). 6.- The problem remains whatever output character set is choosen. 7.- The problem remains if I use a spanish version of FM to export the data. Has anyone else experienced this issue? (it is very easy to reproduce). Unfortunately, I cannot use other export format, because several databases are feed using CSV files and reprogramming this could be a nightmare. Any ideas? Thank you

Thank you. That seems to solve the issue (at least in my trials). If it is a bug (as it seems it is), I hope the guys from FM solve it in the next patch. Thanks again