How to extract information from more PDF files?

Hi,

I’m building a RAG agent and creating a workflow that lets me upload multiple PDFs to Google Drive (in this case, I tested with two). In the RAG pipeline, the data should be split into chunks and tagged with the correct file name from Google Drive. The idea is that it should work regardless of how many PDFs I add (sometimes one, sometimes several).

The first three nodes in the pipeline work as expected. The issue is with the “Extract from file” node: it extracts two items, but they both end up with the same title. When I use the output of that node later (in the Code node and Supabase), only one PDF’s information gets saved.

What could be the problem?

This is the output from the “Extract from file” node:

2 items
numrender
12
info
PDFFormatVersion
1.4
EncryptFilterName
[null]
IsLinearized
true
IsXFAPresent
false
IsCollectionPresent
false
CreationDate
D:20250616111702+02’00’
Creator
Adobe InDesign 20.3 (Windows)
Producer
Adobe PDF Library 17.0
Trapped
metadata
xmp:createdate
2025-06-16T11:17:02+02:00
xmp:metadatadate
2025-06-16T11:17:07+02:00
xmpmm:instanceid
uuid:15c2a323-e7c0-452c-bc7b-6692a4397eb9
xmpmm:originaldocumentid
xmp.did:54DE385ADD01E01191D7A8908C81849F
xmpmm:documentid
xmp.id:387dad5d-041d-684d-ab20-5855b6451724
xmpmm:renditionclass
proof:pdf
dc:format
application/pdf
pdf:producer
Adobe PDF Library 17.0
pdf:trapped
False
text
text[0]
Breakthrough\nengineering for\na better world\nClimate\nControl\nIMI TA\nTA-Nano, TA-Nano Plus\nCombined control & balancing valves for small\nterminal units\nPressure independent balancing and control valve\n(PIBCV)
text[1]
IMI TA / Control valves / TA-Nano, TA-Nano Plus\n2\nTA-Nano, TA-Nano Plus\nThe pressure-independent balancing and control valve TA-Nano\nensures optimum performance over a long life. Adjustable maximum\nflow enables design flow and eliminates overflows for accurate\nhydronic control. The TA-Nano Plus together with our balancing\ninstruments enables advanced measuring and diagnostics.\nKey features\nSmallest PIBCV in the market fitting\nwithin the most constrained area\nSlim and compact shape simplifies\ninstallation.\nPrecise hydronic balancing\nSmoothly adjustable setting of max. flow\nprevents over flow th…
text[2]
3\nIMI TA / Control valves / TA-Nano, TA-Nano Plus\nTemperature:\nMax. working temperature: 120 °C\nMin. working temperature: –10 °C\nNote: If the medium temperature is\nbelow 2 °C, then ice forming on the\nspindle must be prevented. Therefore\nvalves should be insulated with vapor\ntight insulation (stem extension can\nbe used). IMI valves were tested for\nperformance and durability with mono-\nethylene as well as mono-propylene\nglycol up to a concentration of 57%.\nMedia:\nWater or neutral fluids, water-glycol\nmixtures (0-57%).\nLift:\n4 mm\nLeakage rate:\nTight sealing (Class VI according to\nEN 60534-4).\nCha…
text[3]
IMI TA / Control valves / TA-Nano, TA-Nano Plus\n4\nCorrection factors\nThe flow calculations are valid for water (+20°C). For other liquids with approximately the same viscosity as water\n(≤20 cSt = 3°E = 100S.U.), it is only necessary to compensate for the specific density. However, at low temperatures, the\nviscosity increases and laminar flow may occur in the valves. This causes a flow deviation that increases with small valves, low\nsettings and low differential pressures. Correction for this deviation can be made with the software HySelect or directly in our\nbalancing instruments.\nNoise\nIn ord…
text[4]
5\nIMI TA / Control valves / TA-Nano, TA-Nano Plus\nSizing\n1. Choose the smallest valve size that can obtain the design flow with some safety margin, see “q\nmax \nvalues”. The setting should\nbe as open as possible.\n2. Check that the available ΔpV is within the working range ΔpV\nmin \n(according to DN) - 600 kPa.\nq\nmax \nvalues\nLow flow (LF) Normal flow (NF) High flow (HF)\nPosition\n1 2 3 4 5 6 7 8 9 10\nDN 10 NF 19 38 57 76 95 114 133 152 171 190\nDN 15 LF 29 58 87 116 145 174 203 232 261 290\nDN 15 NF 55 110 165 220 275 330 385 440 495 550\nDN 15 HF 105 210 315 420 525 630 735 840 945 1050\nDN 20 NF 110…
text[5]
IMI TA / Control valves / TA-Nano, TA-Nano Plus\n6\nInstallation\nApplication example\nFlow direction\nInstallation of actuator\nNote: Free space is required above the actuator for easy\nmounting/dismounting.\nTA-Slider 160\nH\nEMO T/TM\nH\nTA-TRI\nH\nDN 10-25 122 122 106\nTA-Nano + TA-Slider / EMO T/TM ⁄ TA-TRI\nTA-Nano TA-Nano\nTA-Nano TA-Nano\nTA-Nano TA-Nano\nTA-Nano Plus TA-Nano Plus\nTA-Nano Plus TA-Nano Plus\nTA-Nano Plus TA-Nano Plus\nH \nH\nH\n30 \n30\n30\nIP54 IP54 IP 54 IP54
text[6]
7\nIMI TA / Control valves / TA-Nano, TA-Nano Plus\nOperating function\nStandard / Plus versions\nSetting\n1. Turn the setting wheel to desired value, e.g. 5.0.\nShut-off\n1. Turn the setting wheel clockwise to 0.\nPlus version\nMeasuring q\n1. Remove the installed actuator.\n2. Connect the IMI TA balancing instrument to the measuring\npoints.\n3. Input the valve type, size and setting and the actual flow is\ndisplayed.\nMeasuring ΔH\n1. Remove any actuator.\n2. Close the valve according to “Shut-off”.\n3. Connect IMI TA balancing instrument to the measuring\npoints and measure.\nImportant! Reopen the valve to pr…
text[7]
IMI TA / Control valves / TA-Nano, TA-Nano Plus\n8\nArticles - Standard, without measuring points\nExternal thread\nThreads according to ISO 228.\nDN d L H q\nmax\n[l/h]\nKg EAN Article No\n10 NF G1/2 65 68 190 0,31 5902276824005 322213-00110\n15 LF G3/4 65 68 290 0,35 5902276824012 322213-00015\n15 NF G3/4 65 68 550 0,35 5902276824029 322213-00115\n15 HF G3/4 65 68 1050 0,35 5902276824036 322213-00215\n20 NF G1 75 68 1100 0,38 5902276824043 322213-00120\n20 HF G1 75 68 (1600) 0,38 5902276824050 322213-00220 Launch Sep -25\n25 NF G1 1/4 82 68 (2200) 5902276824067 322213-00125 Launch Sep -25\nInternal thread\nT…
text[8]
9\nIMI TA / Control valves / TA-Nano, TA-Nano Plus\nArticles - Plus, with measuring points\nExternal thread\nThreads according to ISO 228.\nDN d L H H1 q\nmax\n[l/h]\nKg EAN Article No\n10 NF G1/2 65 68 72 190 0,43 5902276824074 322213-10110 Launch Sep -25\n15 LF G3/4 65 68 72 290 0,47 5902276824081 322213-10015 Launch Sep -25\n15 NF G3/4 65 68 72 550 0,47 5902276824098 322213-10115 Launch Sep -25\n15 HF G3/4 65 68 72 1050 0,47 5902276824104 322213-10215 Launch Sep -25\n20 NF G1 75 68 72 1100 0,51 5902276824111 322213-10120 Launch Sep -25\n20 HF G1 75 68 72 (1600) 0,51 5902276824128 322213-10220 Launch Sep …
text[9]
IMI TA / Control valves / TA-Nano, TA-Nano Plus\n10\nConnections\nWith internal thread\nThreads according to ISO 228. Thread length according to ISO 7-1.\nSwivelling nut.\nBrass\nFor DN D D1 L* EAN Article No\n10 G1/2 G3/8 29,5 5902276820014 52 009-810\n10 G1/2 G1/2 34,5 5902276820021 52 009-910\n15 G3/4 G1/2 31,5 5902276820038 52 009-815\n15 G3/4 G3/4 36,5 5902276820045 52 009-915\n20 G1 G3/4 33,5 5902276820052 52 009-820\n20 G1 G1 39,5 5902276820069 52 009-920\n25 G1 1/4 G1 39 5902276820076 52 009-825\n25 G1 1/4 G1 1/4 43 5902276820083 52 009-925\nWith external thread\nThreads according to ISO 7-1.\nSwivellin…
text[10]
11\nIMI TA / Control valves / TA-Nano, TA-Nano Plus\nConnection with smooth end\nFor connection with press coupling.\nSwivelling nut.\nBrass/AMETAL\n®\nFor DN D Pipe Ø L* EAN Article No\n10 G1/2 12 35 7318793810502 52 009-312\n15 G3/4 15 39 7318793810601 52 009-315\n20 G1 18 44 7318793810700 52 009-318\n20 G1 22 48 7318793810809 52 009-322\n25 G1 1/4 28 53 7318793810908 52 009-328\nCompression connection\nSupport bushes shall be used, for more information see catalogue leaflet FPL.\nShould not be used with PEX pipes.\nBrass/AMETAL\n®\nChrome plated\nFor DN D Pipe Ø L** EAN Article No\n10 G1/2 10 17 7318793620101 …
text[11]
IMI TA / Control valves / TA-Nano, TA-Nano Plus\nAccessories\nProtection cap\nFor TA-Nano, TA-COMPACT-P/-DP, TA-Modulator (DN 10-20), TBV-C/-CM.\nColour EAN Article No\nRed 7318793961105 52 143-100\nTamper proof cover\nSet containing plastic cover and locking ring for valves with connection M30x1.5 to thermostatic\nhead/actuator.\nPrevents manipulation of setting.\nEAN Article No\n7318794030206 52 164-100\nSpindle extension\nRecommended together with the insulation to minimize the risk of condensation at the valve-\nactuator interface.\nM30x1,5.\nType L EAN Article No\nPlastic, black 30 4024052165018 2002-30.7…
version
5.3.31

Information on your n8n setup

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:

Hi,

As you can see on the attached image, the “Download file” does get both of the file names.

Is it possible to keep unique file names with “Extract from file” node or I need to adjust the code to get it from the “Download file” node?

Thanks.

P.S. This is what I get as outcome from my current “Code” node (2 items but one fileName and one fullText - don’t mind the pageNumber).

Should I put a for loop node or something on after “Extract from file” node, because I only see information about one of the PDFs (image below it the output from it)? Does it only go file by file?

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.