LinkedIn
Twitter
1 ACCEPTED SOLUTION
For the second half of the fix, I wrote a couple of Javascript functions to parse the JSON data in the collated file and extract only the row data for each table that had sync data. I ran these functions in the Chrome browser console (because you can execute JS from there) and then copied the console log data into json2table.com to get it in tabular format. Then I pasted them into a spreadsheet to then clean for entry into the DB.
Interesting note: It turns out there were only 200 files (200 rows) in the Recovery File, not 444. I wonder if AppSheet caps the Recovery Data process at 200 sync items?
@praveen are you aware of any limit on the recovery file creation?
I suppose I could have written a few more lines of JS to just generate the html table, or even just run the whole bit in AppScript and used JDBC to go right into the DB… Next time 
Here’s the code for anyone who has the same issue.
function getTables (data) {
var tableNames = [];
for (i = 0; i < data.length; i++) {
tableNames.push(data[i].tableName);
};
var distictTableNames = [...new Set(tableNames)];
console.log(distictTableNames);
return distictTableNames;
};
function extractRows (data) {
var tables = getTables(data);
var numTables = tables.length;
var betterData = [];
for (j = 0; j < numTables; j++) {
console.log("entering for loop")
var thisTable = tables[j];
console.log("looping through this tables rows: "+thisTable );
var filterData = data.filter(table => table.tableName === thisTable);
betterData.push([]);
console.log(`${thisTable} has ${filterData.length} rows`);
for (i = 0; i < filterData.length; i++) {
betterData[j].push(filterData[i].row);
};
console.log(JSON.stringify(betterData[j]));
};
console.log(`Looped through data and extracted rows for ${numTables}: ${tables}`);
return true;
};
I got the first part figured out!
I joined them into one file with a Command Prompt command.
Now to parse the 13,000 line JSON file 
Thanks for the suggestions, everyone!
For the second half of the fix, I wrote a couple of Javascript functions to parse the JSON data in the collated file and extract only the row data for each table that had sync data. I ran these functions in the Chrome browser console (because you can execute JS from there) and then copied the console log data into json2table.com to get it in tabular format. Then I pasted them into a spreadsheet to then clean for entry into the DB.
Interesting note: It turns out there were only 200 files (200 rows) in the Recovery File, not 444. I wonder if AppSheet caps the Recovery Data process at 200 sync items?
@praveen are you aware of any limit on the recovery file creation?
I suppose I could have written a few more lines of JS to just generate the html table, or even just run the whole bit in AppScript and used JDBC to go right into the DB… Next time
Here’s the code for anyone who has the same issue.
function getTables (data) {
var tableNames = [];
for (i = 0; i < data.length; i++) {
tableNames.push(data[i].tableName);
};
var distictTableNames = [...new Set(tableNames)];
console.log(distictTableNames);
return distictTableNames;
};
function extractRows (data) {
var tables = getTables(data);
var numTables = tables.length;
var betterData = [];
for (j = 0; j < numTables; j++) {
console.log("entering for loop")
var thisTable = tables[j];
console.log("looping through this tables rows: "+thisTable );
var filterData = data.filter(table => table.tableName === thisTable);
betterData.push([]);
console.log(`${thisTable} has ${filterData.length} rows`);
for (i = 0; i < filterData.length; i++) {
betterData[j].push(filterData[i].row);
};
console.log(JSON.stringify(betterData[j]));
};
console.log(`Looped through data and extracted rows for ${numTables}: ${tables}`);
return true;
};
This is lame… Like it would be cool if we had some appsheet developer tools to be able to handle this recovery…
Recovery used to be epic. You’d turn on recover mode and the system would spit the records into the data source - it was up to you (the developer) to be there to catch and fix the records as they came in.
Then it turned into a nightmare. So much so that when a client and I faced trying to recover a couple hundred (not even 400 so Kudos to @Stefan_Quartemont), after spending an hour trying to find a smooth way to recover the records we abandoned it and had people re-enter things. 
There needs to be a better way.
Hi Stefan, thank you for creating this tool. I tried it this weekend but for some reason I can't get the final step to work.
I've successfully uploaded the files, but the "Process" button doesn't become grey and clickable as per step no 7 in your instructions.
Your advice will be much appreciated.
