Ocean West

If you import the child table first and the data set already has a foreign key - or natural or unique values that can be used as a temporary foreign key then you can import the child and loop thru the records to create a "parent" record. In some cases the data set (child has enough fields that concatenated together can substitute temporarily for a foreign key ) to also make finding matching values you could use GetContainerAttribute ( Lower ( fieldA & fieldB & fieldC ) ; "md5" ) to use as a way to create a parent record. Once the parent record is created (establishing a UUID) do a replace on the child table for the permanent foreign key then adjust the relationship.

ExecuteSQL ( "Select Price from CountryToysPrice WHERE ToyID = ? and country = ?" ; "" ; "" ; Table::GlobalToySelect ; Table::GlobalCountrySelect ) ExecuteSQL ( "Select Price from CountryToysPrice WHERE ToyName = ? and country = ?" ; "" ; "" ; Table::GlobalToySelect ; Table::GlobalCountrySelect ) You would need to have A global field that has values to select from.

For best performance you would want to go to the lowest level of the BOM and cache the summary total in to the parent as trying to aggregate data depending on how large the data set is can be taxing on performance. Alternatively you could create aggregate summary tables to store this data there are many examples how to provide this option over at filemakerhacks.com https://filemakerhacks.com/2018/05/30/fast-summaries-re-revisited/

tap on image navigate to another layout only showing the graphic turn off zoom lock while on layout - use back button in top/bottom navigation to return previous layout setting zoom lock

I am not certain why you'd need to do a search and replace externally why not have the data formatted out as needed? I agree with @comment that there could be much more easier way to structure the data reduce the number of fields. Alternatively build a fully formed HTML Template and put in place holders @@FIELDNAME@@ then output a calculation that smartly replaces the placeholders with real data. Even look at using BaseElements Plugin to create files encoded as needed and write out lines to the file as you are looping thru records. vs using the Export records option.

Reworked the script for you Script.pdfReliaprint1.fmp12.zip

Please provide some sample data with matching field names and table names.

What is the value of LH_101? is it a return separated list such as dog¶cat¶mouse¶horse¶fish

The field name and the table name must match exactly - since it's a string if you change the table name or field name it will fail. Search.fmp12

you said you have 2000 records of which there may be only 10 unique SearchKey items we need to reduce that list of 2000 down to that list so when the script is looping it only has to iterate (search) for the 10 items producing the 10 exports. by using the ExecuteSQL function you can derive a return list of values.

Since you are on FM 16 there isn't an option to 'create" directories - so you will have to manually make sure there is a folder called OUTPUT on the desktop otherwise it will fail. If you need to stay in FM16 you can use BaseElements Plugin or an AppleScript to create a the OUTPUT directory. Also on FM16 you don't have the UniqueValues function, a custom function may be required to obtain a unique list of values ( line 7 is commented out ) Calculation for lIne 7: ExecuteSQL ( "SELECT Distinct SearchKey from PVC_Banners" ; "" ; "" )

Hopefully this will work.... in the search i added "==" so that it searches on exact terms. Search.fmp12

@comment clever use of SerialIncrement - aways forget that one.

Let ( [ from = GetAsTimestamp ( DateOfClass & " " & StartTime ) ; to = GetAsTimeStamp ( DateOfClass & " " & EndTime ) ]; DateOfClass & " " & GetAsTime ( from ) & "-" & GetAsTime( to ) )

It may be possible with MirrorSync