I'm setting up a Google Cloud Functions (GCF) function that gets triggered often enough that there are multiple instances running at the same time.
I am getting errors from a readStream the source file of the stream does not exist, but at this point in my program I've actually just created it.
I've made sure the file exists before the start of the stream by console.log()-ing the file JSON, so the file does actually exist. I've also made sure that the file I'm trying to access has finished being written by a previous stream with an await, but no dice.
EDIT: The code now contains the entire script. The section that seems to be throwing the error is the function columnDelete().
var parse = require('fast-csv');
var Storage = require('@google-cloud/storage');
var Transform = require('readable-stream').Transform;
var storage = new Storage();
var bucket = storage.bucket('<BUCKET>');
const DMSs = ['PBS','CDK','One_Eighty','InfoBahn'];
class DeleteColumns extends Transform{
constructor(){
super({objectMode:true})
}
_transform(row, enc, done){
//create an array 2 elements shorter than received
let newRow = new Array(row.length - 2);
//write all data but the first two columns
for(let i = 0; i < newRow.length; i++){
newRow[i] = row[i+2];
}
this.push(newRow.toString() + '\n');
done();
}
}
function rename(file, originalFile, DMS){
return new Promise((resolve, reject) => {
var dealer;
var date;
var header = true;
var parser = parse({delimiter : ",", quote:'\\'});
//for each row of data
var stream = originalFile.createReadStream();
stream.pipe(parser)
.on('data', (row)=>{
//if this is the first line do nothing
if(header){
header = false;
}
//otherwise record the contents of the first two columns and then destroy the stream
else {
dealer = row[0].toString().replace('"', '').replace('"', '');
date = row[1].toString().replace('"', '').replace('"', '');
stream.end();
}
})
.on('finish', function(){
var newName = dealer + ' ' + date + '_' + DMS + 'temp.csv';
//if this was not triggered by the renaming of a file
if(!file.name.includes(dealer)&&!file.name.includes(':')){
console.log('Renamed ' + file.name);
originalFile.copy(newName);
originalFile.copy(newName.replace('temp',''));
}else{
newName = 'Not Renamed';
console.log('Oops, triggered by the rename');
}
resolve(newName);
});
});
}
function columnDelete(fileName){
return new Promise((resolve, reject) =>{
console.log('Deleting Columns...');
console.log(bucket.file(fileName));
var parser = parse({delimiter : ",", quote:'\\'});
var del = new DeleteColumns();
var temp = bucket.file(fileName);
var final = bucket.file(fileName.replace('temp', ''));
//for each row of data
temp.createReadStream()
//parse the csv
.pipe(parser)
//delete first two columns
.pipe(del)
//write to new file
.pipe(final.createWriteStream()
.on('finish', function(){
console.log('Columns Deleted');
temp.delete();
resolve();
})
);
});
}
exports.triggerRename = async(data, context) => {
var DMS = 'Triple';
var file = data;
//if not a temporary file
if(!file.name.includes('temp')){
//create a new File object from the name of the data passed
const originalFile = bucket.file(file.name);
//identify which database this data is from
DMSs.forEach(function(database){
if(file.name.includes(database)){
DMS = database;
}
});
//rename the file
var tempName = await rename(file, originalFile, DMS);
//if it was renamed, delete the extra columns
if (!tempName.includes('Not Renamed')){
await columnDelete(tempName);
}
} else if(file.name.includes('undefined')){
console.log(file.name + ' is invalid. Deleted.');
bucket.file(file.name).delete();
}
else {
console.log( file.name + ' is a temporary file. Did not rename.');
}
};
What I expect to be output is as below:
Deleting Columns...
Columns Deleted
Nice and simple, letting us know when it has started and finished.
However, I get this instead:
Deleting Columns...
ApiError: No such object: <file> at at Object.parseHttpRespMessage(......)
finished with status: 'crash'
Which is not wanted for obvious reasons. My next thought is to make sure that the file hasn't been deleted by another instance of the script midway through, but to do that I would have to check to see if the file is being used by another stream, which is, to my knowledge, not possible.
Any ideas out there?