CSV Data

We provide the relevant part *.EDM4EIC.root data converted to the CSV format

• The CVS files are located in csv folder in the campaign directory. See DATA ACCESS
• File names start the same as source edm4eic.root file and correspond to each other. E.g. k_lambda_5x41_5000evt_001.*
• We also provide .csv.zip - zipped versions. Pandas can work with such files out of the box
• Access to the CSV and .csv.zip files is the same. See DATA ACCESS page
• CSV table names are embedded in extension before .csv , e.g. *.mcdis.csv, *.mcpart_lambda.csv
• Column names are listed in the first line of the file (standard for CSV)

Example file names:

# Original file
k_lambda_5x41_5000evt_001.edm4eic.root

# Related CSV-s
k_lambda_5x41_5000evt_001.mcdis.csv
k_lambda_5x41_5000evt_001.mcpart_lambda.csv

# Zi

All scripts that make EDM4HEP to CSV conversion are located at csv_convert dir.

Table definitions

For analyzing data, we can work with multiple CSV files that contain related information. The files are linked relationally. The first columns of a CSV table is always a primary key (e.g. event number). Or a composite key (e.g. event number + particle index). For example, all data related to e.g. k_lambda_5x41_5000evt_001.* will refer the same events.

These CSV files are essentially database tables, and understanding this relationship helps us organize and analyze data more effectively.

With python and pandas it is easy to organize them joined tables like MCvsReconstructed events

mc_dis

• Files: *.mc_dis.csv
• Conversion script: csv_convert/csv_mc_dis.cxx

True event level values that come from the event generator. event - evnet id in file, the rest of the names correspond to table: mc-variables

Columns:

event
alphas
mx2
nu
p_rt
pdrest
pperps
pperpz
q2
s_e,s_q
tempvar
tprime
tspectator
twopdotk
twopdotq
w
x_d
xbj
y_d
yplus

reco_dis

• Files: *.reco_dis.csv
• Conversion script: csv_convert/csv_reco_dis.cxx

Reconstructed (and true MC) event kinematic parameters including the reconstructed scattered electron information, beam particles, Lambda particles, and various t-value calculations.

EICRecon provides several algorithms calculating the DIS kinematics. We save them all to CSV. E.g. jb_q2 corresponds to Q2 obtained by Jacquet-Blondel method and electron_q2 corresponds to scattered electron method.

DIS Kinematics Columns

Prefixes - EDM4EIC Collection name:

1. da - "InclusiveKinematicsDA"
2. esigma - "InclusiveKinematicsESigma"
3. electron - "InclusiveKinematicsElectron"
4. jb - "InclusiveKinematicsJB"
5. ml - "InclusiveKinematicsML"
6. sigma - "InclusiveKinematicsSigma"
7. mc - True MC values from event parameters

For each kinematic method (da, esigma, electron, jb, ml, sigma, mc), variables are saved like:

1. {}_x - Bjorken x
2. {}_q2 - Q² [GeV²]
3. {}_y - Inelasticity y
4. {}_nu - Energy transfer ν [GeV]
5. {}_w - Invariant mass W [GeV]

T-value Columns

The script calculates several t-values (momentum transfer squared) using different beam configurations:

1. mc_true_t - True t-value from MC event parameters (dis_tspectator)
2. mc_lam_tb_t - t calculated using MC Lambda and true beam proton
3. mc_lam_exp_t - t calculated using MC Lambda and experimental beam proton
4. ff_lam_tb_t - t calculated using far-forward reconstructed Lambda and true beam
5. ff_lam_exp_t - t calculated using far-forward reconstructed Lambda and experimental beam

Important Physics Note:

• True beam uses the actual MC beam proton momentum from the simulation
• Experimental beam approximates what we would know in a real experiment:
  • Detects the beam mode (41, 100, 130, or 275 GeV) from the true momentum
  • Applies crossing angles: 25 mrad horizontal, 100 μrad vertical
  • This mimics experimental conditions where we don't know the exact beam momentum

Scattered Electron Columns

For the reconstructed scattered electron (from the Electron method):

1. elec_id - Particle index in ReconstructedParticles collection
2. elec_energy - Total energy [GeV]
3. elec_px - Momentum x-component [GeV/c]
4. elec_py - Momentum y-component [GeV/c]
5. elec_pz - Momentum z-component [GeV/c]
6. elec_ref_x - Reference point x-coordinate
7. elec_ref_y - Reference point y-coordinate
8. elec_ref_z - Reference point z-coordinate
9. elec_pid_goodness - Particle ID quality metric
10. elec_type - Reconstruction type flag
11. elec_n_clusters - Number of associated clusters
12. elec_n_tracks - Number of associated tracks
13. elec_n_particles - Number of daughter particles
14. elec_n_particle_ids - Number of particle ID objects

MC Scattered Electron Momentum

1. mc_elec_px - MC truth scattered electron px [GeV/c]
2. mc_elec_py - MC truth scattered electron py [GeV/c]
3. mc_elec_pz - MC truth scattered electron pz [GeV/c]

Lambda Momentum Columns

MC truth Lambda:

1. mc_lam_px - MC Lambda px [GeV/c]
2. mc_lam_py - MC Lambda py [GeV/c]
3. mc_lam_pz - MC Lambda pz [GeV/c]

Far-forward reconstructed Lambda:

1. ff_lam_px - Far-forward Lambda px [GeV/c]
2. ff_lam_py - Far-forward Lambda py [GeV/c]
3. ff_lam_pz - Far-forward Lambda pz [GeV/c]

Beam Particle Momentum Columns

MC beam proton:

1. mc_beam_prot_px - Beam proton px [GeV/c]
2. mc_beam_prot_py - Beam proton py [GeV/c]
3. mc_beam_prot_pz - Beam proton pz [GeV/c]

MC beam electron:

1. mc_beam_elec_px - Beam electron px [GeV/c]
2. mc_beam_elec_py - Beam electron py [GeV/c]
3. mc_beam_elec_pz - Beam electron pz [GeV/c]

event is the first column = event number.

So the complete column list is:

event,
da_x,da_q2,da_y,da_nu,da_w,
esigma_x,esigma_q2,esigma_y,esigma_nu,esigma_w,
electron_x,electron_q2,electron_y,electron_nu,electron_w,
jb_x,jb_q2,jb_y,jb_nu,jb_w,
ml_x,ml_q2,ml_y,ml_nu,ml_w,
sigma_x,sigma_q2,sigma_y,sigma_nu,sigma_w,
mc_x,mc_q2,mc_y,mc_nu,mc_w,
mc_true_t,mc_lam_tb_t,mc_lam_exp_t,ff_lam_tb_t,ff_lam_exp_t,
elec_id,elec_energy,elec_px,elec_py,elec_pz,elec_ref_x,elec_ref_y,elec_ref_z,elec_pid_goodness,elec_type,elec_n_clusters,elec_n_tracks,elec_n_particles,elec_n_particle_ids,
mc_elec_px,mc_elec_py,mc_elec_pz,
mc_lam_px,mc_lam_py,mc_lam_pz,
ff_lam_px,ff_lam_py,ff_lam_pz,
mc_beam_prot_px,mc_beam_prot_py,mc_beam_prot_pz,
mc_beam_elec_px,mc_beam_elec_py,mc_beam_elec_pz

Notes:

• The electron particle information is only available when the Electron method successfully reconstructs the scattered electron
• If particles are not found/reconstructed, their columns will contain null values
• T-values are calculated as t = (p1 - p2)² using 4-vectors
• The experimental beam approximation is crucial for understanding systematic uncertainties in real experiments

mcpart_lambda

• Files: *.mcpart_lambda.csv
• Conversion script: csv_convert/csv_mcpart_lambda.cxx

Full MC particles information for a lambda decays chain by using MCParticles EDM4EIC table. MCParticles has relations like daughters and parents. Those relations are flattened by lambda decay. The columns represent possible lambda decays are grouped by particles:

Prefixes (each has the same parameters after)

1. lam - Λ
2. prot - p (if pπ- decay or nulls)
3. pimin - π- (if pπ- decay or nulls)
4. neut - Neutron (if n π0 decay)
5. pizero - pi0 - (if n π0 decay)
6. gamone - γ one from π0 decay (if pi0 decays)
7. gamtwo - γ two from π0 decay (if pi0 decays)

For each particle prefix, the next columns are saved:

1. {0}_id - id - particle index in MCParticles table
2. {0}_pdg - pdg - particle PDG
3. {0}_gen - gen - Generator Status (1 stable... probably)
4. {0}_sim - sim - Simulation Status (by Geant4)
5. {0}_px - px - Momentum
6. {0}_py - py
7. {0}_pz - pz
8. {0}_vx - vx - Origin vertex information
9. {0}_vy - vy
10. {0}_vz - vz
11. {0}_epx - epx - End Point (decay, or out of detector)
12. {0}_epy - epy
13. {0}_epz - epz
14. {0}_time - time - Time of origin
15. {0}_nd - nd - Number of daughters
16. {0}_np - np - Number of parents

In addition to the particle blocks, each row starts with two event/decay columns:

1. event — event number (primary key).
2. lam_is_first — 1 if this Λ is the first Λ encountered in the event (the generated spectator Λ), else 0. One row per Λ is written, so files can contain multiple rows per event if more than one Λ appears.
3. lam_decay — decay-channel code (see table below). This is the shared decay code used by all three Λ CSVs (mcpart_lambda, acceptance_npi0, acceptance_ppim).

Code	Meaning	Daughter rule
0	Not decayed	daughters.size() == 0
1	Λ → p π⁻	2 daughters: PDG 2212 + PDG -211
2	Λ → n π⁰	2 daughters: PDG 2112 + PDG 111
3	Shower	daughters.size() > 2 (any)
4	Only p	1 daughter, PDG 2212
5	Only π⁺	1 daughter, PDG 211
6	Only n	1 daughter, PDG 2112
7	Only π⁰	1 daughter, PDG 111
8	Other	anything not matching above

So in the end the columns are:

event,lam_is_first,lam_decay,
lam_id,lam_pdg,lam_gen,lam_sim,lam_px,lam_py,lam_pz,lam_vx,lam_vy,lam_vz,lam_epx,lam_epy,lam_epz,lam_time,lam_nd,lam_np,
prot_id,prot_pdg,prot_gen,prot_sim,prot_px,prot_py,prot_pz,prot_vx,prot_vy,prot_vz,prot_epx,prot_epy,prot_epz,prot_time,prot_nd,prot_np,
pimin_id,pimin_pdg,pimin_gen,pimin_sim,pimin_px,pimin_py,pimin_pz,pimin_vx,pimin_vy,pimin_vz,pimin_epx,pimin_epy,pimin_epz,pimin_time,pimin_nd,pimin_np,
neut_id,neut_pdg,neut_gen,neut_sim,neut_px,neut_py,neut_pz,neut_vx,neut_vy,neut_vz,neut_epx,neut_epy,neut_epz,neut_time,neut_nd,neut_np,
pizero_id,pizero_pdg,pizero_gen,pizero_sim,pizero_px,pizero_py,pizero_pz,pizero_vx,pizero_vy,pizero_vz,pizero_epx,pizero_epy,pizero_epz,pizero_time,pizero_nd,pizero_np,
gamone_id,gamone_pdg,gamone_gen,gamone_sim,gamone_px,gamone_py,gamone_pz,gamone_vx,gamone_vy,gamone_vz,gamone_epx,gamone_epy,gamone_epz,gamone_time,gamone_nd,gamone_np,
gamtwo_id,gamtwo_pdg,gamtwo_gen,gamtwo_sim,gamtwo_px,gamtwo_py,gamtwo_pz,gamtwo_vx,gamtwo_vy,gamtwo_vz,gamtwo_epx,gamtwo_epy,gamtwo_epz,gamtwo_time,gamtwo_nd,gamtwo_np

Notes:

• Particles may not be decayed. E.g. Lambda may just go outside of detector designated volume, in this case lam_nd - Number of daughters will be 0 and the rest of columns will be null.
• Particle-block columns that don't apply to the observed decay channel are left empty (e.g. neut_*, pizero_*, gamone_*, gamtwo_* are empty when lam_decay == 1).

acceptance_npi0

• Files: *.acceptance_npi0.csv
• Conversion script: csv_convert/csv_edm4hep_acceptance_npi0.cxx

Same per-Λ rows as mcpart_lambda, plus per-particle MC-truth acceptance flags for the Λ → n π⁰ → n γ γ chain. One row per first Λ per event. Detection flags are 1 only when the Λ decayed via n+π⁰ and the π⁰ produced two observable photons; otherwise they are 0. A flag is 1 iff the particle contributed to at least one SimCalorimeterHit in that detector collection.

Column prefix (identical to mcpart_lambda):

event,lam_is_first,lam_decay,
<lam_*>, <prot_*>, <pimin_*>, <neut_*>, <pizero_*>, <gamone_*>, <gamtwo_*>

Followed by detection flags (column name = <particle>_<EDM4hep collection name>):

Neutron in HCALs:

1. neut_HcalFarForwardZDCHits
2. neut_HcalEndcapPInsertHits
3. neut_LFHCALHits

Gamma-one in ECALs:

4. gamone_EcalFarForwardZDCHits
5. gamone_B0ECalHits
6. gamone_EcalEndcapPHits
7. gamone_EcalEndcapPInsertHits

Gamma-two in ECALs (same four):

8. gamtwo_EcalFarForwardZDCHits
9. gamtwo_B0ECalHits
10. gamtwo_EcalEndcapPHits
11. gamtwo_EcalEndcapPInsertHits

Full header ordering:

event,lam_is_first,lam_decay,
lam_id,…,lam_np, prot_id,…,prot_np, pimin_id,…,pimin_np,
neut_id,…,neut_np, pizero_id,…,pizero_np, gamone_id,…,gamone_np, gamtwo_id,…,gamtwo_np,
neut_HcalFarForwardZDCHits, neut_HcalEndcapPInsertHits, neut_LFHCALHits,
gamone_EcalFarForwardZDCHits, gamtwo_EcalFarForwardZDCHits,
gamone_B0ECalHits,             gamtwo_B0ECalHits,
gamone_EcalEndcapPHits,        gamtwo_EcalEndcapPHits,
gamone_EcalEndcapPInsertHits,  gamtwo_EcalEndcapPInsertHits

Notes:

• Detection flags are 0 for all non-n+π⁰ decays (nothing is looked up for those rows).
• has hit means the MCParticle appears in at least one SimCalorimeterHit::getContributions() of the collection.

acceptance_ppim

• Files: *.acceptance_ppim.csv
• Conversion script: csv_convert/csv_edm4hep_acceptance_ppim.cxx

Per-event acceptance flags for the Λ → p π⁻ channel. Only events whose first-reachable Λ is a p π⁻ decay produce a row. lam_decay is therefore always 1 in this file (it is kept for schema consistency with the other two Λ CSVs).

Each particle (prot, pimin) gets a per-collection 0/1 flag for every EDM4hep tracker and calorimeter collection. Column names use the full EDM4hep collection name: <particle>_<CollectionName>.

Tracker collections (17) checked for both prot and pimin:

B0TrackerHits, BackwardMPGDEndcapHits, DIRCBarHits, DRICHHits,
ForwardMPGDEndcapHits, ForwardOffMTrackerHits, ForwardRomanPotHits,
LumiSpecTrackerHits, MPGDBarrelHits, OuterMPGDBarrelHits,
RICHEndcapNHits, SiBarrelHits, TOFBarrelHits, TOFEndcapHits,
TaggerTrackerHits, TrackerEndcapHits, VertexBarrelHits

Calorimeter collections (7) checked for both prot and pimin:

EcalFarForwardZDCHits, B0ECalHits, EcalEndcapPHits, EcalEndcapPInsertHits,
HcalFarForwardZDCHits, HcalEndcapPInsertHits, LFHCALHits

Flag rule: 1 if the particle has at least one hit (tracker: SimTrackerHit::getParticle() matches; calo: any contribution matches), else 0.

Full header ordering:

event,lam_is_first,lam_decay,
lam_id,…,lam_np, prot_id,…,prot_np, pimin_id,…,pimin_np,
prot_<tracker-1>,…,prot_<tracker-17>,
prot_<calo-1>,…,prot_<calo-7>,
pimin_<tracker-1>,…,pimin_<tracker-17>,
pimin_<calo-1>,…,pimin_<calo-7>

Companion hit-detail CSVs

The ppim script additionally writes two long-format per-hit CSVs, one row per hit:

• *.acceptance_ppim_prot_hits.csv — detailed hits belonging to the proton candidate
• *.acceptance_ppim_pimin_hits.csv — detailed hits belonging to the π⁻ candidate

Columns (identical for both):

Column	Description
event	Event number (join key to the main ppim CSV).
lam_id	MCParticle index of the parent Λ (join key to lam_id in main CSV).
detector	EDM4hep collection name (B0TrackerHits, HcalFarForwardZDCHits, …).
hit_id	SimHit::getObjectID().index.
x,y,z	Hit position [mm].
eDep	Energy deposit [GeV] (tracker: getEDep(), calo: getEnergy()).
time	Hit time [ns]. For calo hits taken from the first matching contribution.
pathLength	Tracker getPathLength(). Always 0 for calorimeter rows.

Join pattern:

import pandas as pd
main = pd.read_csv("run.acceptance_ppim.csv")
prot_hits = pd.read_csv("run.acceptance_ppim_prot_hits.csv")
merged = prot_hits.merge(main, on=["event"], suffixes=("_hit", ""))

reco_ff_lambdas

• Files: *.reco_ff_lambdas_ngamgam.csv
• Conversion script: csv_convert/csv_reco_ff_lambda.cxx

Reconstructed Lambda particles and their decay products from the far-forward Zero Degree Calorimeter (ZDC), specifically for the decay channel Λ → n + π⁰ → n + γ + γ. This table uses the ReconstructedFarForwardZDCLambdas collection from EDM4EIC and flattens the decay hierarchy similar to mcpart_lambda.

The columns are grouped by particles in the decay chain:

Prefixes (each has the same parameters after):

1. lam - Λ (Lambda baryon)
2. neut - Neutron from Λ decay
3. gam1 - First γ from π⁰ decay
4. gam2 - Second γ from π⁰ decay

For each particle prefix, the following columns are saved:

1. {0}_id - id - particle index in ReconstructedParticles collection
2. {0}_pdg - pdg - particle PDG code
3. {0}_charge - charge - electric charge
4. {0}_energy - energy - total energy [GeV]
5. {0}_mass - mass - invariant mass [GeV/c²]
6. {0}_px - px - momentum x-component [GeV/c]
7. {0}_py - py - momentum y-component [GeV/c]
8. {0}_pz - pz - momentum z-component [GeV/c]
9. {0}_ref_x - ref_x - reference point x-coordinate
10. {0}_ref_y - ref_y - reference point y-coordinate
11. {0}_ref_z - ref_z - reference point z-coordinate
12. {0}_pid_goodness - pid_goodness - particle ID quality metric
13. {0}_type - type - reconstruction type flag
14. {0}_n_clusters - n_clusters - number of associated clusters
15. {0}_n_tracks - n_tracks - number of associated tracks
16. {0}_n_particles - n_particles - number of daughter particles
17. {0}_n_particle_ids - n_particle_ids - number of particle ID objects
18. {0}_cov_xx - cov_xx - covariance matrix element
19. {0}_cov_xy - cov_xy - covariance matrix element
20. {0}_cov_xz - cov_xz - covariance matrix element
21. {0}_cov_yy - cov_yy - covariance matrix element
22. {0}_cov_yz - cov_yz - covariance matrix element
23. {0}_cov_zz - cov_zz - covariance matrix element
24. {0}_cov_xt - cov_xt - covariance matrix element
25. {0}_cov_yt - cov_yt - covariance matrix element
26. {0}_cov_zt - cov_zt - covariance matrix element
27. {0}_cov_tt - cov_tt - covariance matrix element

The complete column list is:

event,
lam_id,lam_pdg,lam_charge,lam_energy,lam_mass,lam_px,lam_py,lam_pz,lam_ref_x,lam_ref_y,lam_ref_z,lam_pid_goodness,lam_type,lam_n_clusters,lam_n_tracks,lam_n_particles,lam_n_particle_ids,lam_cov_xx,lam_cov_xy,lam_cov_xz,lam_cov_yy,lam_cov_yz,lam_cov_zz,lam_cov_xt,lam_cov_yt,lam_cov_zt,lam_cov_tt,
neut_id,neut_pdg,neut_charge,neut_energy,neut_mass,neut_px,neut_py,neut_pz,neut_ref_x,neut_ref_y,neut_ref_z,neut_pid_goodness,neut_type,neut_n_clusters,neut_n_tracks,neut_n_particles,neut_n_particle_ids,neut_cov_xx,neut_cov_xy,neut_cov_xz,neut_cov_yy,neut_cov_yz,neut_cov_zz,neut_cov_xt,neut_cov_yt,neut_cov_zt,neut_cov_tt,
gam1_id,gam1_pdg,gam1_charge,gam1_energy,gam1_mass,gam1_px,gam1_py,gam1_pz,gam1_ref_x,gam1_ref_y,gam1_ref_z,gam1_pid_goodness,gam1_type,gam1_n_clusters,gam1_n_tracks,gam1_n_particles,gam1_n_particle_ids,gam1_cov_xx,gam1_cov_xy,gam1_cov_xz,gam1_cov_yy,gam1_cov_yz,gam1_cov_zz,gam1_cov_xt,gam1_cov_yt,gam1_cov_zt,gam1_cov_tt,
gam2_id,gam2_pdg,gam2_charge,gam2_energy,gam2_mass,gam2_px,gam2_py,gam2_pz,gam2_ref_x,gam2_ref_y,gam2_ref_z,gam2_pid_goodness,gam2_type,gam2_n_clusters,gam2_n_tracks,gam2_n_particles,gam2_n_particle_ids,gam2_cov_xx,gam2_cov_xy,gam2_cov_xz,gam2_cov_yy,gam2_cov_yz,gam2_cov_zz,gam2_cov_xt,gam2_cov_yt,gam2_cov_zt,gam2_cov_tt

Notes:

• ZDC reconstructed lambdas look only Lambda decays (Λ → n + π⁰ → n + γ + γ channel)
• If a particle is not reconstructed or missing, its columns will contain null values
• The n_particles field for the Lambda indicates the number of reconstructed daughter particles

ppim_combinatorics

• Files: *.ppim_combinatorics.csv
• Conversion script: csv_convert/csv_edm4hep_combinatorics_ppim.cxx
• Analysis script: csv_convert/analyse_ppim_combinatorics.py

Combinatoric proton + π⁻ candidate pairs for Λ → p + π⁻ searches using far-forward detectors. Each row is one (proton candidate, pion candidate) pair found in an event — events with zero candidates produce no rows.

Candidate selection (MC-truth level):

Candidate	Detector collection	Min hits
Proton	ForwardRomanPotHits	≥ 2
Pion (π⁻)	B0TrackerHits	≥ 3

Pion candidates are additionally checked against B0ECalHits contributions; a flag records whether any calorimeter energy was deposited.

Truth matching: The script locates the primary Λ in each event and stores its true daughter particle IDs in true_prot_id / true_pi_id (both are -1 when no primary Λ decay is found). The is_true_lam flag is 1 when the row's candidate pair exactly matches those true daughters.

Note: the event-ID column is evt (not event as in other tables).

Columns

Event / truth metadata:

1. evt — event number
2. is_true_lam — 1 if this pair is the true p + π⁻ from the primary Λ, else 0
3. true_prot_id — MCParticle index of the true Lambda daughter proton (−1 if none)
4. true_pi_id — MCParticle index of the true Lambda daughter π⁻ (−1 if none)

Pion candidate — MCParticle info (same 16-field schema as other tables, prefix pi):

5–20. pi_id, pi_pdg, pi_gen, pi_sim, pi_px, pi_py, pi_pz, pi_vx, pi_vy, pi_vz, pi_epx, pi_epy, pi_epz, pi_time, pi_nd, pi_np

Pion candidate — B0 tracker hit info:

21. pi_nhits_b0 — number of B0TrackerHits for this candidate
22. pi_first_b0_x — x of first registered B0TrackerHit [mm]
23. pi_first_b0_y — y of first registered B0TrackerHit [mm]
24. pi_first_b0_z — z of first registered B0TrackerHit [mm]

Pion candidate — B0 ECal contribution:

25. pi_ecal_contrib — 1 if candidate left ≥1 hit contribution in B0ECalHits, else 0
26. pi_first_ecal_x — x of first B0ECal cell with contribution [mm] (0 if none)
27. pi_first_ecal_y — y of first B0ECal cell with contribution [mm] (0 if none)
28. pi_first_ecal_z — z of first B0ECal cell with contribution [mm] (0 if none)

Proton candidate — MCParticle info (prefix prot):

29–44. prot_id, prot_pdg, prot_gen, prot_sim, prot_px, prot_py, prot_pz, prot_vx, prot_vy, prot_vz, prot_epx, prot_epy, prot_epz, prot_time, prot_nd, prot_np

Proton candidate — Roman Pot hit info:

45. prot_nhits_rp — number of ForwardRomanPotHits for this candidate
46. prot_first_rp_x — x of first registered ForwardRomanPotHit [mm]
47. prot_first_rp_y — y of first registered ForwardRomanPotHit [mm]
48. prot_first_rp_z — z of first registered ForwardRomanPotHit [mm]

Complete column list:

evt,
is_true_lam,true_prot_id,true_pi_id,
pi_id,pi_pdg,pi_gen,pi_sim,pi_px,pi_py,pi_pz,pi_vx,pi_vy,pi_vz,pi_epx,pi_epy,pi_epz,pi_time,pi_nd,pi_np,
pi_nhits_b0,pi_first_b0_x,pi_first_b0_y,pi_first_b0_z,
pi_ecal_contrib,pi_first_ecal_x,pi_first_ecal_y,pi_first_ecal_z,
prot_id,prot_pdg,prot_gen,prot_sim,prot_px,prot_py,prot_pz,prot_vx,prot_vy,prot_vz,prot_epx,prot_epy,prot_epz,prot_time,prot_nd,prot_np,
prot_nhits_rp,prot_first_rp_x,prot_first_rp_y,prot_first_rp_z

Notes:

• Because rows are only written when at least one proton and one pion candidate exist, events with zero candidates are absent from the file entirely.
• Columns true_prot_id / true_pi_id can be joined against pi_id / prot_id to identify reversed assignments — cases where the true proton reached B0 and the true pion reached the Roman Pots — using analyse_ppim_combinatorics.py.
• When combining multiple files, offset evt the same way as event in other tables (see Combine Multiple Files below).

Combine Multiple Files

When we have multiple CSV files from different runs or datasets, each file starts its event numbering from 0:

File 1: event = [0, 1, 2, 3, 4, ...]
File 2: event = [0, 1, 2, 3, 4, ...]  ← ID Collision!
File 3: event = [0, 1, 2, 3, 4, ...]  ← ID Collision!

Problem: Event 0 from File 1 is completely different from Event 0 from File 2, but they have the same ID if read in pandas directly!

Use functions like this to read multiple files in one DF

import pandas as pd
import glob

def concat_csvs_with_unique_events(files):
    """Load and concatenate CSV files with globally unique event IDs"""
    dfs = []
    offset = 0
    
    for file in files:
        df = pd.read_csv(file)
        df['event'] = df['event'] + offset  # Make IDs globally unique
        offset = df['event'].max() + 1    # Set offset for next file
        dfs.append(df)
    
    return pd.concat(dfs, ignore_index=True)

# Load both tables with unique event IDs
lambda_df = concat_csvs_with_unique_events(sorted(glob.glob("mcpart_lambda*.csv")))
dis_df = concat_csvs_with_unique_events(sorted(glob.glob("dis_parameters*.csv")))

Result: Now we have globally unique event IDs:

File 1: event = [0, 1, 2, 3, 4]
File 2: event = [5, 6, 7, 8, 9]     ← No collision!  
File 3: event = [10, 11, 12, 13, 14] ← No collision!