Latest Contributions by DataOfficer

So after some testing and discussions with an ESRI consultant, we came up with a solution. Sharing here in case anyone else has the same issue. General background: The waypoints (artificial refugia for reptiles in this case) for each site are stored in an external select (external_choices tab of the form), with a SiteName column specifying which waypoints belong to which sites. This enables a choice_filter to be implemented based on the site selected. A separate external csv (in the media folder) stores sites and waypoints in wide format (column 1 = SiteName, column 2 = Waypoint 1 name, column 3 = Waypoint 2 name ..... column n = Waypoint n-1 name) - up to the maximum number of waypoints a site may have (e.g. 100). Process: The user selects the site (or autopopulated from Explorer / Collector using a URL to load the form) Within a repeat: The user selects a waypoint from a select_one dropdown list, with a choice_filter controlling the list based on the site selection. A waypoint label calculate field uses the jr:choice-name calculation to store the label of that waypoint. name: Waypoint_label calculation: jr:choice-name(${Waypoints},'${Waypoints}') Outside of the repeat: A text field uses a join calculation to join the label of each waypoint surveyed so far. name: Waypoints_checked calculation: join(",",${Waypoint_label} This text field shows which waypoints have been checked so far. We created hidden rows up to the maximum number of waypoints at a site (e.g. 100 rows), each with a relevant function and a pulldata calculation to the external csv for each waypoint column (e.g. row 1 of 100 pulls the label of waypoint 1). relevant: not(contains(${Waypoint_checked}, pulldata('Waypoints','Waypoint1','SiteName',${SiteName}))) calculation: pulldata('Waypoints','Waypoint1','SiteName',${SiteName})  For aesthetic reasons, a second set of hidden rows were created up to the maximum number of waypoints (e.g. 100 rows), which is relevant based on the string-length of the corresponding pulldata row above being >0. The calculation field then concatenates the waypoint name with a space. E.g. name: Waypoint_1 relevant: string-length(${Waypoint1})>0 calculation: concat(${Waypoint1}," ") This then allows concatenation below without additional commas appearing if a site did not have the maximum number of waypoints, compared to if we just concatenated the outputs from the pulldata calculation (step 3.2). Another row is used to concatenate all of the rows that were concatenated in step 3.3 name: WaypointsNotChecked calculation: concat(${Waypoint_1},${Waypoint_2},${Waypoint_3},${Waypoint_4}, The output of step 3.4 is then presented as a note: <center><b>Warning!</b> The following waypoints for ${Site_label} have not been checked: <b></center> <center><b><font color="red">${WaypointsNotChecked}</font></b></center> <center>If these are missing, please add records for these waypoints and select <i>missing</i> from <i>status</i>. <b>Only proceed if you are finished</b></center>          The warning notes are dependent on a relevant calculation which hides them if the string length of WaypointsNotChecked is 0.  So in summary, the user gets two outputs at the end of the survey form: a list of waypoints checked so far (Waypoints_checked), and a warning showing the waypoints which have not been checked. Thanks Doug Browning and others for their input. ... View more